Web Services Making Software Coexist?

jgeelan writes "Despite the competitive uproar, coexistence of J2EE and .NET will be the norm and most sophisticated IT organizations will deploy on both development platforms. Who says so? No less an authority than the CTO of J2EE powerhouse BEA Systems, Scott Dietzen, writing in this month's Web Services Journal. Dietzen acknowledges that an ongoing conflict is in progress between Java and C# and between J2EE and the .NET server family and is refeshingly honest, admitting that "there is some truth to the 'write once, test everywhere' complaint against Java." His overall conclusion: ".NET is finding a sweet spot for programmed user interfaces, while J2EE continues to enjoy its sweet spot for server-side applications." Unusual honesty by someone so highly placed. Isn't this just what the software industry needs more of, in these increasingly interoperable times?"

Comments

i want neither...

[packeteer]

... its all about .mac

Re:i want neither...

[whatparadox]

I want them all. For my mac, my wintel, my hey-the-video-card-with-this-release Redhat. As a SysAdmin for a small company I don't wan to worry if the 'new thing' will run on the Marketing Depts OS9 macs, the COO's OSX laptop, or the linux that runs the POS in the pizza joints.

Re:i want neither...

[dolithe]

I am impresed, computers and human-computer
interaction have come a long way. People find
computers and their language useful, hard,
entertaining, humorous and even sexy!!! I wonder if the
oposite holds as well?

Milan
http://www.dolithe.com

I can see it now!

[Izanagi]

I think an 800 pound gorilla is about to get a caffine buzz from drinking java!

Honesty or idiocy?

[AJWM]

".NET is finding a sweet spot for programmed user interfaces,"

On a single platform, perhaps. It's true enough that early editions of Java's Swing weren't the swiftest UIs on the block, Swing has to contend with being platform independant. How well does a .NET UI run on a Solaris workstation?

The UI for most server-side applications is probably HTML, anyway, so I'm not sure what his point was. I suspect BEA is just making nice noises toward .NET to gain a foot in the door of MS-only shops, although it is certainly true that at the SOAP level .NET and J2EE could interact.

We'll see how long that lasts when/if .NET market share catches up with J2EE. (Embrace and extend, anyone?)

Re:Honesty or idiocy?

[TummyX]

On a single platform, perhaps. It's true enough that early editions of Java's Swing weren't the swiftest UIs on the block, Swing has to contend with being platform independant. How well does a .NET UI run on a Solaris workstation?

The UI for most server-side applications is probably HTML, anyway, so I'm not sure what his point was.


That is the point. C#, ASP.NET and VS.NET are the perfect combination if you want to make HTML based UIs. You just drag a button onto a webpage, double click and write the event handling code in C#. The C# is compiled into a server side component along with some ASPX pages. The UI is completely HTML based and cross platform. The HTML generated can be transformed to work on all types of browsers (utilising the features of each). Anything that can't be done client side is done on the server side.

You pretty much write HTMl pages the same way you would write a standard GUI application and .NET takes care of the rest. Its a great way to program HTML UIs. And I think that was part of his point.

Re:Honesty or idiocy?

[wadetemp]

How well does a .NET UI run on a Solaris workstation?

Actually you answered your own question. It runs fine, if its an ASP.NET UI. Working on HTML/JavaScript based UI is nearly the same process as developing a client side UI in .NET, given the event-based programming model and similar UI classes for both. Granted, you can't yet run a pure Windows.Forms application on Solaris, but you didn't ask "how well does a .NET Windows Forms UI run on a Solaris workstation," did you? :)

Re:Honesty or idiocy?

[bitdamaged]

so true because Solaris is used on soooooooo many workstations, these days its obvious that Dietzen's generalization is obviously wrong, what the hell is he doing as the CTO of BEA anyway.

Jeez isn't it fun being a zealot.

Re:Honesty or idiocy?

[AJWM]

You just drag a button onto a webpage, double click and write the event handling code in C#. The C# is compiled into a server side component along with some ASPX pages.

Okay, I'll grant you that seems like a pretty cool development environment (for that sort of thing). But it isn't .NET specific, there's no fundamental reason why a similar IDE couldn't be done for JSP. Indeed, it probably already has.

Re:Honesty or idiocy?

[AJWM]

you didn't ask "how well does a .NET Windows Forms UI run on a Solaris workstation," did you? :)

Well not specifically, no, but the context was comparing with Java and Swing. Swing is hardly the UI used by JSP or servlets, HTML/JavaScript is.

Kind of a moot question, otherwise, eh?

Re:Honesty or idiocy?

[RevAaron]

I don't know about for Java, but VisualWorks Smalltalk (there's a free, non-commercial version for download for Mac OS, Mac OS X, Linux, Windows, many other Unices) that does something like this. Using one single UI builder, you construct a GUI for the traditional GUI application. From the same spec it generates, a web application and interface. You define callbacks and such just like you would for a regular app, and VisualWave (the web app toolkit) takes care of the rest of it.

VisualWorks has had this ability for quite a while, at least 4 years or so, which was the first time I played with it. I don't doubt that there may have been something before it that did the same thing, but it preceeds .NET and probably even a similar Java tool.

Re:Honesty or idiocy?

[DunbarTheInept]

The programmer's IDE is not a feature of the language nor the .net archetecture. For example, tab-completion of function names is not a feature of Visual Basic. It's a feature of MS Developer Studio.

HTML generated can be transformed to work on all types of browsers (utilising the features of each)

Or it could have just been written to te W3c spec in the first place. No, that wouldn't be the MS way. Instead make something that has to know which browser you are using and change its behaviour accordingly. (Which of course fails utterly to be *forward* compatable with new and improved browsers you haven't heard of yet. I'm sick of websites that tell me I need to "upgrade" to Netscape 4.76 or I.E. 5.0 when I'm running something *newer* that would work if the site would just send me the HTML it would have sent to one of those two browsers.)

Re:Honesty or idiocy?

[The B]

Smalltalk has been phased out by IBM, and the widget elemnts etc are old. Perhaps we could use .NET to fix programmer UI errors like the slashdot messageboard UI ugh!

Re:Honesty or idiocy?

[zacks]

no one here seems to acknowledge macromedia products here, slashdot readers amkes the world seems to be php and mysql based... mysql is not a proper relational database (yet).... just becase something is free doesn't mean it's the best solution... when a developer costs more than one software license per day, the argument looks a little weak..

the real agurment here is how long does it take to develop out of the box a simple or complex application... cold fusion is simple a much faster language, yes it's actually more 4gl than 3gl, but what are u trying to do? pull some data out of the database and display it....

in cf that means 2 tags (opened and closed) and var names and html

<cfquery name="test" datasource="localserver">
select name, email, description
from users
where active = 1
order by name
</cfquery>

<cfoutput query="test">
#name# - #email#
#description#
</cfoutput>

looks quite clean huh?

why not look at at coldfusion mx, a j2ee application server, where the tag based language makes things simple and you can always fall back to java.

The when you want a more complex decent ui than is possible with html, use flash, w3c dom is great but not so prevalent..

and flash makes for a much for reliable user expierence than dhtml

Re:Honesty or idiocy?

[m3573]

Smalltalk has been phased out by IBM, and the widget elemnts etc are old.

Squeak is being developed and it's open source: it looks like a toy but its approach to user interaction using "morphs" is interesting.

Cincom (its smalltalk site slashdotted? :) ) is developing version 7 of visualworks smalltalk, and it will surely be out before i finish checking out all the features of version 5 :)

The widgets are a little dated indeed, but being a very modular system there may be GUI enhancements to download, try looking around.

I'm impressed with some smalltalk environment features, like being able to modify some code (which gets dynamically recompiled, no "build" command) and having the changes reflected on the application while it is running, being able to inspect and extend system classes (smalltalk applications merely extend the system, in fact), being able to save and restore all classes and variables i'm working with in a single image file or in distinct packages.

I'd like to know if there is a java environment with similar features.

Re:Honesty or idiocy?

[Jonboy X]

The whole point of "web services" is that pretty much any two networked devices can communicate. WS just says "Use HTTP. Pretty much everything can support that". It says NOTHING about using HTML! You don't use a web browser to talk to a web service. You use a web service client that can generate and understand the XML (probably) that is flying over the wire (or through the air). If your client uses a GUI to display data to a user, it can use whatever language it wants, or maybe just shove data into a database for examination later.

The motivation behind using HTTP is a dubious one at best. The basic idea is to get around corporate firewalls. This can be liberating, I s'pose, but it's also a security breach. WS is only a big deal 'cuz most PHB's have never written raw socket code, but all of them have used a web browser.

Re:Honesty or idiocy?

[TheAncientHacker]

But, some of us actually want to use features not in all browsers. I don't really want to have to code separate versions of every page for Netscape 1.0 and for Lynx. You see, the world has changed since 1994. Really. It has.

Anyone who's done web pages for real has had to deal with browser compatibility. The ASP.NET approach of emitting different HTML for different browsers makes this a LOT less of a pain.

Re:Honesty or idiocy?

[RevAaron]

IBM's VisualAge for Java has features similar to many Smalltalk environments. Surprise, surprise- it was written in Smalltalk. :)

I personally do development in Smalltalk (using Squeak), and I love it. One of my projects over the last few years was enhancing the IRC client built into Squeak. I was adding the ability to have bots. In a couple days of hacking, I didn't have to restart the application once- all objects of the classes being changed had the updated methods as soon as I saved them. Then, I wasn't a newbie anymore, but it still amazed me then, and still does now. I shudder to think how much time extra it would take to have to recompile and restart after every insigificant change of code.

Aaron

Re:Honesty or idiocy?

[AJWM]

The motivation behind using HTTP is a dubious one at best. The basic idea is to get around corporate firewalls.

While I don't disagree that using HTTP is often dubious, getting around firewalls isn't the only reason. (Heck, a lot of this stuff is on intranets behind the firewall anyway.)

Other reasons include:

• HTTP is a "connectionless" protocol, ie the connection is not held open (well, HTTP 1.1 aside). This means if somebody walks away from his computer in the middle of a session he isn't still tying up a socket and process (and possibly a DB connection) on the server. Which means overall you can get away with a smaller server, resources are shared better.
• HTTP is a pretty simple protocol, that means it's easier to implement and can be squeezed into smaller code or devices. It's also pretty ubiquitous, there are all kinds of libraries to support it. Not true if you're rolling your own protocol. (I know, I've rolled a few of my own when HTTP's statelessness is more trouble than it's worth.)
• HTTP lets you use a browser as a standard and ubiquitous client if you need user interaction (as you point out, the client isn't necessarily a browser.) That makes UI development easy and cross-platform. The only other choices that come to mind for readily and widely available clients are things like NNTP, SMTP, FTP or telnet, none of whice really lend themselves to generalized interactive client/server use (except telnet, if you're into the command line interface). Or X, but then you spend a lot of work on the GUI.

And of course now that it has become a big deal, there are a ton of tools out there to make it simpler to develop and deploy such apps.

Re:Honesty or idiocy?

[Bodrius]

I think Apple's WebObjects did a decent job with Java in that respect.

The problem is that the .NET tools are more polished, and the J2EE tools are lacking. The advantage of Java, right now, is the J2EE de facto standard... it's already doing what the companies want it to do, and they already have an investment in it, and there's no reason not to have the equivalent of a VS.NET for J2EE that works without propietary dependencies.

Re:Honesty or idiocy?

[IamTheRealMike]

You pretty much write HTMl pages the same way you would write a standard GUI application and .NET takes care of the rest. Its a great way to program HTML UIs. And I think that was part of his point.

But surely you get a lot less flexibility? I seem to recall MS tried to rewrite Outlook in DHTML, the project bombed, as HTML (even with loads of IE extensions) just wasn't suitable for user interfaces. That's why Mozilla created XUL (which is). How does it deal with things like snap-off toolbars, popup menus, zigzagged underlines and all the other stuff that's either impossible or extremely hard to do with HTML?

It sounds like you'd end up with an extremely limited and rather slow widget set - a bit like AWT in fact. On the other hand, I haven't tried it, so what do I know?

Re:Honesty or idiocy?

[estoll]

You are right on one level; however, completely ignorant of the big picture. First off, let me state that I am a .NET developer for over a year and, honestly, I believe .NET has great potential. My first impression of ASP.NET and Visual Studio.NET was, well, extremely impressed. The fact the most .NET developers I have talked with (including those with over 10-years web development) think .NET is this magic technology that is going to solve all the world's problems is absolutely rediculous! The second these people get some .NET marketing blown up their ass and a pretty IDE in front of their face, they think the world of cross-browser compatibility, HTML, Javascript, etc has been replaced by a God-like alternative. Suddenly, Microsoft makes everyone think if it doesn't work with .NET, it must be you fault (not using Internet Explorer). Big surprise here. This is the strong arm they have been trying to push for years! Will .NET UI run on a Solaris workstation? Hell no! Will Microsoft target this market? Absolutely not! Do they claim they support alternative browsers for their web applications? Sure; however, that is just marketing BS. You may ask what is this magical .NET's ability to support these "alternative" browsers? Let me say, it is minimal. As far as simple web pages go (simple forms, buttons, hyperlinks) sure, they work the rest; however, when it comes to anything higher-level (i.e. the developer actually needing an understanding of how this shit works in the real world without all the fancy IDE (which in my opinion is something any web developer should understand in the first place (how can you develop a web page and not need to understand that HTTP is a stateless protocol as Microsoft "suggests" is not a necessary concept?))). most .NET developers are completely in the dark and when something doesn't "work" it must be _impossible_ to do. No! Microsoft is not dropping-the-ball on this marketing and releasing .NET so early was not a "misstep"-- Microsoft is dilgently attacking the DEVELOPERS into making them believe that the other alternatives "don't work"! Go look at the newsgroups! Microsoft is blatently telling people that if it doesn't work in .NET, it must be the next guy's fault (however, removing themselves from liability by letting they "MVP's" do their work)! So back to my original statement-- I like .NET; I think it has great potential. But, why the hell do you want to replace something that is already working-- something that is already matured-- and something that has proven it's ability, stability, and efficiency?

Re:Honesty or idiocy?

[drew]

It's possible, (and usually rather simple) to program a web page to take advantage of features not in all web pages without explicitly checking browser versions. The parent's point is that if you make your web page say "requires ie version x", then it's going to break when version x+1 comes out. guaranteed. if you make the browser require a feature in version x, rather than requiring a specific version, then your web page will still work when the next version comes out, assuming the feature hasn't been deprecated.

this is harder to do when you actually want to change the HTML code based on the browesr, but in most cases all you're changing is stylesheets or javascript code, which is much simpler. instead of putting "if (navigator.appVersion == x)" in your code, just do something like "if (function_that_i_want_to_use)"

this way you dont have to go back and change your code whenever a new browser is released, and it has the added side bonus of also being more likely to work in browsers that dont have a dominant marketshare but are in many ways feature compatible (i.e. konqueror and opera)

Re:Honesty or idiocy?

[DunbarTheInept]

I don't really want to have to code separate versions of every page for Netscape 1.0 and for Lynx.

Well, it's a good thing that what I advocated was the exact opposite of that, then.

Anyone who's done web pages for real has had to deal with browser compatibility. The ASP.NET approach of emitting different HTML for different browsers makes this a LOT less of a pain.

But the types of servers you speak of don't check for *features* supported. They check for brand names, and that's the complaint I have. Just because a browser visiting your page is one you didn't test for, that doesn't automatically mean it can't do the features you need. It's that broken mentality that makes new up-and-coming browsers like Opera have to lie about what kind of browser they are in order to use certain sites developed with a mentality similar to yours. The reason they have to lie just to get the HTML shipped to them is that the servers aren't programed with a default case of "just spit out w3c standard stuff and if the browser can't deal with it that's the browser's fault." So places that program only to IE, or only to IE and Netscape continue to believe the lie that those are the only browsers that exist (because people have to browse with the user-agent string set to the lie that they are IE or Netscape), and that's what shows up in all logs then.

Re:Honesty or idiocy?

[DunbarTheInept]

The problem you complain about is smaller that the problem that is generated when servers check for browser type and assume that all browsers not of that type will fail utterly. I'd much rather be fed an incompatable page with sloppy layout than NO HTML AT ALL and an error message that just lies and says my browser doesn't support feature foo when it fact it does.

Re:Honesty or idiocy?

[TheAncientHacker]

Actually, ASP.NET reads the header, looks that up to see capabilities and emits based on capabilities.

Re:Honesty or idiocy?

[DunbarTheInept]

Then explain why when opera lies and pretends to be IE it gets a better page sent to it than when it is honest and calls itself opera.

Re:Honesty or idiocy?

[TheAncientHacker]

Simple. Because the capabilities of specific browsers are kept in a table on the server that has to be maintained for new browsers. If the owner of the server doesn't bother then new entries are unrecognized and default back to a more generic set of capabilities.

More Python Please!

[supton]

Ugh... I'm sick of these industry horses wearing blinders. There are more than two choices, of course. When .NET and J2EE developers finally get sick of watching their compilers sing and dance, I'll still be here adding new features to my Python applications at run time in a much better language. :)

Re:More Python Please!

[AJWM]

Oh, you can make J2EE sing and dance with Python, too, if you want.

You have heard of Jython (nee JPython), haven't you?

(And you can add new features to Java applications at run time too (ClassLoader? JavaBeans? hello?). But of course you knew that. :)

Re:More Python Please!

[RevAaron]

And I'll be building even better applications via a web, telnet, or regular GUI interface at runtime in a much better language, Smalltalk. :)

But yeah, you're very right. There are many options other than .NET and Java. I know there isn't much for Python, but there are a few companies (IBM, Cincom, GemStone) that provide real commercial support for their web app server products.

Re:More Python Please!

[The B]

What good RAD development environement is there for Python other than a text editor

Re:More Python Please!

[TheAncientHacker]

Of course, you could always use python.NET to write your ASP.NET pages if you don't want to learn C#.

http://starship.python.net/crew/mhammond/dotnet/ in dex.html

Re:More Python Please!

[RevAaron]

Heh.

If you look hard enough, you'll find similar arguments against all other programming techniques, languages, and frameworks. When it comes down to it, programming at all is just another waste of time.

There's a lot more to OOP than lies and FUD. If it's too hard for you to grasp, that's fine, you can continue using BASIC, no one will mind.

Re:More Python Please!

[supton]

Boa constructor. Good enough to build real apps quicker than something like VB... Integrates with Zope, has very nice code editing; Toolkit is wxPython, so it is cross-platform to Win/*nix/Mac.

http://boa-constructor.sf.net

Re:More Python Please!

[supton]

Perhaps you have the luxury of creating non-modular software all by yourself in a cave. Even in non-OO languages (esp. C, since there are a lot of C programmers about), people still use design patterns that are OOish or hacks that approximate OO. So don't give me the "OO is lies and FUD" BS... explain WHY you think its all one great big lie.

Re:More Python Please!

[supton]

I won't dis Smalltalk. Have you ever noticed that there are a lot of mutual respect between the Smalltalk and Python communities?

Commercial support for Python app servers: My company uses Zope. There are a few strong players in support of Zope in the app server and content management space, including Zope Corp in the US, as well as a bunch of folks in Europe, Brazil, and other places.

Re:More Python Please!

[Tablizer]

(* Perhaps you have the luxury of creating non-modular software all by yourself in a cave. *)

Oh, so procedural/relational programs are "non modular". I would rather have consistent modules rather than speggetti modules of OOP. Every fricken OO guru designs differently from each other. Picasso Black Art that OO. OO is GOTO's at a larger scale.

Further, "modularity" needs to be relative in many cases. There are multiple division candidates for any non-trivial app, and the smarter OO fans and Stepanov (STL) have reconized this problem. One size and one division dimension does not fit all.

P/r apps better handle relativism IMO by using relational and boolean formulas rather than hard-coded noun structures. It is easier to change a formula than shuffle code around.

Formulas are more abstract and flexible than hard-coded GOF-like structures. Think virtual. OOP is old fashioned hand-built bricks and mortar.

There is no fscken evidence that OO makes for better biz apps, and take OO away until you gets some.

"Got Evidence?"

"No, I use OO"

Re:More Python Please!

[Tablizer]

(* Even in non-OO languages (esp. C, since there are a lot of C programmers about), people still use design patterns that are OOish or hacks that approximate OO. *)

Well, C sucks. Burn C, not all procedural/relational paradigms just because one language sucks eggs. Besides, C is usually not used for custom biz apps, the domain of my primary complaint. (Perhaps OO shines in some domains, however, the industry is shoving it down everybody's throat).

(* So don't give me the "OO is lies and FUD" BS... explain WHY you think its all one great big lie. *)

Because there is NO EVIDENCE that it is better. I want to see actual biz-app code under change-impact-analsys, not cliches and brochure-talk meant for gullible PHB's.

Re:More Python Please!

[Tablizer]

(* If it's too hard for you to grasp, that's fine, you can continue using BASIC, no one will mind. *)

Yet Another False Dichotomy. BASIC is hardly the best that procedural can be.

The hardest thing to grasp about OOP is why the hell people would jump (blindly) into it without any evidence that it is better.

Is Science Truly Dead?

Re:More Python Please!

[RevAaron]

Then you can use Pascal, judging by your comments, you're a big fan. You're allowed C if you really want to impress your friends.

What is wrong about using OOP? I like it. It works great for me. I find I can organize code much more logically. Also, sending messages to entities is a more natural way of getting something done, for me. \

You remind me of the atheist who thinks every x-ian to be some preaching asshole. I'm not interested in "converting" you to the "one true way" of OOP, so why do the same to me?

What has science to do with this? Liking OOP rather than procedural programming is a matter of preference, not science. I probably could out-code you if there was some rule about OOP; what that make OOP better? You could probably outcode me if we were doing procedural style. Does that make PP better? Nope.

Re:More Python Please!

[RevAaron]

No need to dis Smalltalk, I wasn't (meaning) to dis Python. It's too bad there is no Python IDE akin to what you get even with a free Smalltalk system, like Squeak. The first language that I really god excited about was Python, but when I discovered Squeak, it marked the end of my Python days. While the language has a lot of appealing aspects to me, the environment was a big thing too.

Oh yes, Zope, good fun there. Zope is interesting- they're basically attempting to recreate a Smalltalk environment for Python, but in a much bigger- in both good and bad.

Re:More Python Please!

[Tablizer]

(* Then you can use Pascal, judging by your comments, you're a big fan. *)

I tend to prefer dynamic languages these days.

(* What is wrong about using OOP? I like it. It works great for me. I find I can organize code much more logically. *)

What *exactly* does that mean? Is this something objective, or just a "feel"?

I realize that for *some* people OO may model the way they personally think, but I don't like other's preferences forced down my throat.

(* I'm not interested in "converting" you to the "one true way" of OOP, so why do the same to me? *)

I am sick of seeing the industry making *everything* OO. Somebody has to stop the nonsense.

I just ask for balance. Until X is objectively proven better than Y, or shown that most programmers prefer X, then X and Y should be given equal treatment.

Simple balance.

Re:More Python Please!

[Zeinfeld]

No need to dis Smalltalk, I wasn't (meaning) to dis Python. It's too bad there is no Python IDE akin to what you get even with a free Smalltalk system, like Squeak.

I believe that there is a Python .NET in the works. The Visual Studio .NET IDE basically has every good feature from pretty much every good IDE. It is the first I have found with a superset of the Genera functionality - and packaged a heck of a lot better than Symbolics ever did.

Re:More Python Please!

[ProfKyne]

So, then, if you don't program in C and you don't program in BASIC, what non-OO language do you program in? Surely you can't be talking about Perl scripts?

I've witnessed your condemnation of OO languages here on slashdot for a while now, but I haven't actually seen you do more than say "Show me the evidence that OO is better". Ironically, in many cases no one is saying that OO is better. They're simply describing what they like. So I'm curious (and I'm genuinely curious, not attacking you), what kind of non-OO code do you write, and in what language? You show me why procedural is better.

Re:More Python Please!

[Tablizer]

(* Ironically, in many cases no one is saying that OO is better. They're simply describing what they like. *)

The literature is full of braggings about OOP. Sure, not everybody thinks it is objectively better, but those are not my complaint anyhow.

(* what kind of non-OO code do you write, and in what language? *)

There is no one procedural language that has all the features I want. Progress in procedural features pretty much stopped in the late 80's as vendors and compiler makers focused on OOP instead. Thus, the best features never got a chance to consolidate before the era of CPU-driven worries dwindled sufficiently.

(* You show me why procedural is better. *)

I *don't* claim it to be objectively better. Just not objectively worse.

But a generally summary of why I think it is better (for me) is that it does not attempt to use code structures to model entities (nouns) and noun/entity relationships. I relies on the database for those, and (good) databases make virtual, local, and relativistic "noun views" easier than OOP. OOP tends to hard-wire them into code, which is more rigid than using relational and boolean formulas IMO. Half of OO GOF is about hand-indexing stuff it feels to me. That is busywork (and clutter) I would rather slave off to the database engines.

Most OO fans don't "get" databases IMO.

Re:More Python Please!

[ProfKyne]

it does not attempt to use code structures to model entities (nouns) and noun/entity relationships. I relies on the database for those, and (good) databases make virtual, local, and relativistic "noun views" easier than OOP. OOP tends to hard-wire them into code, which is more rigid than using relational and boolean formulas IMO.

If you can write code that manages data exclusively via the database, such that it never needs to manifest in the code itself (as an object or, entity, or whatever), then why not just create an instance of an object to represent that record? It doesn't have to be a big, bloated class definition, just a simple thing that you can then use as a reference in your code. Then, if you find that you need to add some kind of feature or behavior to objects of this class, isn't it easier to add it to the class definition (so that it is available throughout the code, wherever the object instance appears) than to have to write a function or some kind of behavior into the application code itself?

I understand what you're saying, and as a back-end web programmer I also kept all of my entity information in the database. But when I needed a new behavior available, I found that it was easier to provide this as a method of the class rather than try to fit it into the procedural parts of my code. Plus, I could always write a new class to aggregate other classes if I needed features of an already-existing class but wanted to expand on those (or I could just use inheritance, but this is more limiting).

At some point, the data needs to be used directly by the code, so flexible data structures such as objects come in handy (rather than cumbersome things like hashes-of-hashes). That's my opinion.

(PS: thanks for responding and not flaming, I am genuinely interested in your perspective, even if I don't agree)

Re:More Python Please!

[Tablizer]

(* then why not just create an instance of an object to represent that record? *)

I find that dictionaries (associative arrays) work just fine. Some languages even use a dot instead of (or in addition to) square brackets.

(* Then, if you find that you need to add some kind of feature or behavior to objects of this class, isn't it easier to add it to the class definition (so that it is available throughout the code, wherever the object instance appears) *)

How is that better than a function?

recordHandle.newBehavior() // OOP

newBehavior(recordHandle) // Function call

In some languages, you can even do this:

Execute(recordHandle["newBehavior"])

essentially using a dictionary entry as a method. However, in practice I don't see a lot of need for such.

(* But when I needed a new behavior available, I found that it was easier to provide this as a method of the class rather than try to fit it into the procedural parts of my code. *)

I would have to see a specific example. Where exactly does the procedural version get harder?

(* Plus, I could always write a new class to aggregate other classes if I needed features of an already-existing class but wanted to expand on those *)

How is this different from a subroutine calling other subroutines?

(* At some point, the data needs to be used directly by the code, so flexible data structures such as objects come in handy (rather than cumbersome things like hashes-of-hashes). *)

I am not sure what you have encountered. Hashes of hashes is usually a sign that you need to design another table.

I used to use a lot of XBase. XBase was not perfect, I would note, but it made creating and using tables so fricken easy that I never wanted to touch data arrays again. Unfortunately, most existing stuff chose to farm that off to objects rather than improving lite-duty table manipulation. OOP killed table innovation implementation.

Microsoft is starting to bring virtual and local table engines back into its products, to their credit. However, their API's are a little clunky IMO. You can wrap them with something simpler, but its strong-typing push complicates much of it IMO.

Re:More Python Please!

[jfx32]

>>How is that better than a function?

>>recordHandle.newBehavior() // OOP

>>newBehavior(recordHandle) // Function call

Namespaces for one. Namespace pollution can occur very easily on a large project with multiple developers. Coupling the data and functionality also allows you to be sure that your data is in a consistent state. OOP is not a holy grail, but can be useful (esp. for things like user interfaces).

Re:More Python Please!

[jfx32]

Here are a few: Boa Constructor, Black Adder, Visual Python , Python Win, PythonWorks Pro, Wing IDE, and Komodo. There are some others, check out editing python source.

Re:More Python Please!

[Tablizer]

(* Namespaces for one. Namespace pollution can occur very easily on a large project with multiple developers..... *)

Well, instead of x.y(), you can call it x_y().

Generally if a routine is meant for only a given module, then you don't really have to worry about overlap in many languages. If by chance you call a global routine with the same name as a local one, then some scope or module qualifier may be warrented.

If it is "general", then you have to be a little more careful. But, subroutine overlap is not necessarily more problematic than class name overlap. (These kind of things are very language-specific, I must note, and there are different solutions, just as there are different "module" solutions in OOP languages.)

Besides, are you suggesting that OOP is better *only* because it allegedly has less name-space overlap problems? If that is the best you can claim about OOP, then why the rush to OO?

Usually people will brag about its "encapsulation" or "abstraction" or "modeling" power, not namespaces. (Although those other guys never justify their preference in black and white. It is just a feeling to them.)

Re:More Python Please!

[jfx32]

(*Well, instead of x.y(), you can call it x_y().*)

You are correct, however I have worked on a project where this was the naming convention (module_function()), and it never failed that someone would break the convention.

(*Besides, are you suggesting that OOP is better *only* because it allegedly has less name-space overlap problems? If that is the best you can claim about OOP, then why the rush to OO?*)

Actually, I mentioned keeping the data in a consistent state. Encapsulation is usually better with an OO language. If your data members are private they are only accessible by the methods of your class. Again this could be achieved by convention with procedural programming, if everyone follows the convention. You could also create a convention for an object heirarchy (like gtk), but it is usually easier if it's part of the language.

(*Usually people will brag about its "encapsulation" or "abstraction" or "modeling" power, not namespaces. (Although those other guys never justify their preference in black and white. It is just a feeling to them.)*)

None of those concepts are impossible with a procedural language, but are sometimes made easier by an OO language, since it is a part of the language.

Re:More Python Please!

[Tablizer]

(* You are correct, however I have worked on a project where this was the naming convention (module_function()), and it never failed that someone would break the convention. *)

But you can also use other approahces if needed. By not hard-wiring it as a requirement, if you need to do something different, you can. For example, commonly-used routines should not be required to use such a convention to keep the code simpler. (The name/expression length should be inversely proportional to its usage frequency.)

Things get messy in OO also if people don't follow conventions. For example, you can't polymorph (substitute) if a programmer did not use the same method names and/or signatures for similar things.

(* Actually, I mentioned keeping the data in a consistent state. Encapsulation is usually better with an OO language. If your data members are private they are only accessible by the methods of your class. Again this could be achieved by convention with procedural programming, if everyone follows the convention. *)

That is malarky. In practice classes don't *own* data, but are merely proxies. There is nothing built into the OO paradigm to make sure only *one* class accesses a given non-language-bound data item. (And if you add that enforcement, it could be added for routines and/or modules also.)

Even if you could completely "wrap" data, then you end up reinventing a lot of data-base like operations in your class interface. Better to factor all that to the database rather than reinvent it over and over for each class IMO. That just creates OO bloat.

(* None of those concepts are impossible with a procedural language, but are sometimes made easier by an OO language, since it is a part of the language. *)

I will beleive it when I see it. Often the conditions assumed for being "easier" or "more change-friendly" are unrealistic. OO doctrine has over-hilighted certain change patterns, making people blind to the change patterns that don't favor OO. It is like TV commercials that pick on say yellow teeth. Before you never noticed it, but after many repeats, your mind focuses in on that (alleged) shortcomming, ignoring others.

For example, teeth whitening toothpaste has more abrasives in it. In the long run it may damage the outer shell of your teeth. But, you are not aware of that problem because you are sold on the idea that yellow teeth are the main evil.

This is what OO doctrine has appeared to do IMO.

So, OO fans hard-wire noun models into code structure and reinvent database operations time and time again and never realize how wasteful and bloaty and maintenance-unfreindly it is because they are "encapsulating" or "polymorphing" like a good little programmer should, according to the passages in the ScrOOPtures.

Re:More Python Please!

[ProfKyne]

I find that dictionaries (associative arrays) work just fine. Some languages even use a dot instead of (or in addition to) square brackets.

Yes, JavaScript has a similar feature.

(* Then, if you find that you need to add some kind of feature or behavior to objects of this class, isn't it easier to add it to the class definition (so that it is available throughout the code, wherever the object instance appears) *)

How is that better than a function?

Well, for one thing it doesn't clutter the global namespace with yet another function name. (I don't mean "true" global, but it will need to be global to any place where you wish to use this function.)

recordHandle.newBehavior() // OOP

newBehavior(recordHandle) // Function call

If you get into more complex data structures, then it can become tedious to keep track of where all of this data is stored: "which array is holding data X, where is data Y coming from". I can see arguments that this isn't that important, from the standpoint of someone who's really really organized but I guess I'm not -- I like being able to refer to an object's attributes from within its methods without having to pass them in as parameters, or (gack) using global variables.

OOP killed table innovation implementation.

I can understand what you mean, even though I like OOP. But, then, that doesn't mean that people such as yourself can't continue to develop (and promote) procedural programming methodologies and technologies. Which, I suppose, is exactly what you're doing. In any event, thanks at least for clarifying your perspective to me somewhat via this thread, although I may not agree.

Since when have .Net/C# proven themselves?

[burgburgburg]

I'm sorry, but for all the "honesty" of this article, Dietzen is trying to act as if there is an equivalence of evolution between .Net and J2EE, and Java and C#. Despite the huge Microsoft push, there is still an enormous amount of development of .Net and C# that still needs to come, and a significant amount of widespread acceptance/usage before they should be included in the same article.

Java and J2EE are realities that real world developers have been working on for years. Years.

Have C# and .Net even been fully documented yet? How many months do you have to have worked with them to be considered old dog on the development team? Have they been in the wild, real world tested yet? Are they really anything other then Microsoft marketing concepts that they're giving the full court press to?

Re:Since when have .Net/C# proven themselves?

[wadetemp]

How many months do you have to have worked with them to be considered old dog on the development team?

18? It's not like it just came out yesterday, the beta has been available for at least a year and a half. Wrox put out books on programming with the beta languages and the beta framework.

Re:Since when have .Net/C# proven themselves?

[PyroPunk]

The consulting firm I work at has deployed a Point Of Sale system running on the .Net framework; it's currently running in 4 restaurants. The touchscreen GUI was written in C#, as was the business and data tiers. We use .Net remoting to communicate between the touchscreen and the server. Just started development using the .Net compact framework to let the waiters use Pocket PC devices when taking orders. I've been developing in .Net since June of last year; deployed two intranet sites and an e-commerce site since then. Have been developing in Java since '97. I personally don't consider one better than the other; but then again I'm not out there fighting a religious battle against Microsoft either.

Re:Since when have .Net/C# proven themselves?

[embarcadero]

Call me a cynical deconstructionist, but I think it's more important to realize what BEA is trying to do with comments like this.

Clearly, BEA needs to grow over the next decade, and has to be very careful about the inevitable market power that Microsoft will have with .Net. Whether or not you you like it versus J2EE, Microsoft has the deep pockets to basically buy a market position for .Net over the next five years. BEA is king of the hill today, in a Java dominated market. As .Net grows, there will be a new niche opening. That might be filled by Microsoft itself, but it will be just as attractive for a more enterprise-oriented new company to fill if BEA doesn't.

BEA isn't stupid. It realizes it has to be ready as a potential Microsoft partner, or it risks limiting its dominant position in the application servers and related software market to the Java side of the world. That's why it's being so "honest" about things like this.

Re:Since when have .Net/C# proven themselves?

[AJWM]

18 months is just about long enough to get the experience needed to shake out the first set of design bugs. I vaguely recall it was about that time frame to go from Java 1.0 to 1.1, which saw some significant improvements. But there were still a few more more signficant changes from 1.1 to 1.2 (aka Java2).

Granted, with .NET Microsoft has the advantage of learning some lessons that Java pioneered, but it's still a Microsoft 1.0 product. There'll be shakeouts yet.

Re:Since when have .Net/C# proven themselves?

[f00zbll]

It's good to know someone sees C# and java for what it is, A Tool! IMO, it's good that those devices are switching to standard protocols, but it's hardly new. Wireless point of sales systems for restaurants have existed since early 90's. I know for a fact because in southern California, I went to a restaurant where a friend worked that had a wireless sytem. My friend would take the order and it would be sent wireless to a terminal for the cooks to see.

Having standard ways to do remote communication is good, but all the bloated hype from all the various camps does a lot of dis-service. As far as I'm concerned, all the big tech companies are guilty of it, so it's not like MS or Sun are worse than others. The real benefit for businesses with technologies like C#, Java, .NET and web services is standard protocols and frameworks. The chances of a company finding skilled replacement when staff leave is more likely than a non-standard proprietary system. For me, the hassle of marketoid gibberish is worth it to move towards more standard and flexible ways of developing software.

Re:Since when have .Net/C# proven themselves?

[gaspyy]

Are your questions rhetorical, you do you actually want some answers?

My company was involved since last year up until July in big project for a major health insurance company and everything was done with .net

So yes, it's in the wild, worked with and tested.

I'm not saying it's a wonderful technology and - tipical for MS - we had many problems in the early stages. Still, considering the client's needs, it was the most effective solution.

Interoperability's a bitch...

[camusatan]

Probably the biggest complaint that one could have against web services is the difficulty involved in getting separate ones to interoperate.

Imagine you have a web-based e-mail system. And a web-based word-document reader. One written in J2EE and one written in .NET. Click on a word attachment in the e-mail program, any guess as to whether or not it will open in your word-document reader? Answer is nope!

That people are just now figuring out that demanding that all things be written in One True Language to be hosted on One True Platform from One Designated Service Provider is kinda sad - doesn't that sound restrictive?

Of course, my commentary on this issue isn't coming from nowhere - I wrote a service designed to fix the problem, at its core. But just because I'm biased doesn't mean I'm wrong.

Re:Interoperability's a bitch...

[mccalli]

Imagine you have a web-based e-mail system. And a web-based word-document reader. One written in J2EE and one written in .NET. Click on a word attachment in the e-mail program, any guess as to whether or not it will open in your word-document reader? Answer is nope!

Err...yes, of course it will if you've set the MIME type correctly.

Cheers,
Ian

Re:Interoperability's a bitch...

[camusatan]

If you have actual physical real software running on your local machine that your web browser can call out to (e.g, via MIME-type to program association), then yes, you're absolutely right.

But that's not what I'm talking about, here. How do you get one server-based app (which is viewed through a web-browser) to take advantage of some other server-based app (not on the same server - hell, not even hosted from the same place)? The important trick to note here is - without using any client software? How do you do it?

That is what my stuff purports to solve.

Honesty doesn't count for much these days.

[SlashChick]

It's an unfortunate truth that saying something honest will get you much fewer eyeballs. Slashdot even perpetuates this by posting purposely controversial viewpoints. I mean, which is more controversial, "Linux and Windows both have their place" or

"Linux on the desktop is dead! Windows on the server is dead!"

The fact is that the "x and y both have their place" articles, although honest, are much less likely to get published, because media outlets know that less controversial opinions just aren't talked about as much.

This thread will be lucky to receive 200 comments. Had the article submitted been "Java is dying! .Net is taking over!" it probably would have had 800-1000 comments. It's sad, but true.

Having said that, I use PHP. Stop by #php on irc.openprojects.net! </plug> ;)

Re:Honesty doesn't count for much these days.

[God! Awful]

"Today was much like any other day" is an honest and concise summary of my life, and yet it leads to frequent mundane conversations with my parents, who think I am secretive. The fact is, an article that says "x and y have their place" simply isn't news, just like you wouldn't post a story "cold fusion not discovered today" or "freebsd not hacked today."

I also disagree with your thread acitivity. This story would no doubt get more responses if it wasn't posted late in the evening.

-a

Re:Honesty doesn't count for much these days.

[jgeelan]

"This thread will be lucky to receive 200 comments. Had the article submitted been 'Java is dying! .Net is taking over!' it probably would have had 800-1000 comments. It's sad, but true." Well, this thread is already disproving you. Developers are becoming increasingly savvy about the need to be industry-aware as well as technology-aware. So in fact, "x and y both have their place" articles (as you call them) are going to more and more, not less, frequent. No longer is everything just a zero-sum game.

Re:OMFG WHEEEEE

[sbrown123]

".NET is finding a sweet spot for programmed user interfaces, while J2EE continues to enjoy its sweet spot for server-side applications."

So its basically business as usual?

What is a website?

[ambisinistral]

I think a lot of IT Departments haven't dialed into what a website really is yet. They view a website as a large monolithic application.

Management is trying to apply the tools of application design to web design (and by design I mean the structure/function of it -- not the eye candy). Management is also trying to reduce the toolkit design down to one integrated package. Of course, it is also very much in the interest of the large software companies to promote that same concept... fufill all your needs from us, etc., etc. To me, reduced to its simplest, a website is nothing more than navigation scheme for a large information space. That information is plugged in through a large number of essentially descrete applications. As long as the front end of the site looks the same, it doesn't matter how that content is being handled on the backend. I read a paper a long time ago where the author argues that the seperation of the interface from the content was one of the misunderstood strengths of web delivered content. That since virtually any backend application could be plugged into the interface, it could effectively break the monopoly that venders have on IT shops. Oracle gets to be to expensive? Port your data to another DB and change the ODBC drivers and say adios to 'em. For now the lure of the monolithic approach has traction with IT managers used to that approach. It will cost them in flexibility down the road. Just like the big iron mainframe folks had to adopt to the world of client-server networked PCs, the client-server folks are going to have to adapt to the logic of web delivered apps IMO.

Re:What is a website?

[the eric conspiracy]

They view a website as a large monolithic application.

Applications haven't been "monolithic" for a long time. MVC had it's origins in application design, and anything decent these days is designed with the same 3-tier to n-tier approach that a web site would have.

It's all in the economics - you have to be able to update one part of the application without the changes propagating throughout the whole application.

Re:What is a website?

[ambisinistral]

Applications haven't been "monolithic" for a long time.

I agree 100% However, now try telling an IT manager that the intranet they're developing isn't a monolithic application. ;-)

In fact your comment, "It's all in the economics - you have to be able to update one part of the application without the changes propagating throughout the whole application" is exactly what I was trying to get across. In the old client-server model when you changed an application its interface went with it. Because the interface and the content delivery were so tightly bound together changing the application meant the interface, out of necessaity, changed also. So all of the training and interoperability issues reared their head.

With web-based delivery that isn't the case. Designed properly, the interface and content of a web site are segregated from each other. Designed properly, you can switch out the backend without the user ever even noticing. Or, you can redesign the front end look and feel without ever having to touch the backend apps. Rhe people who keep mushing content and interface together in one entity are going to have maintenance problems, the shops that don't are going to spend much less time spinning the wheels corrected self-inflicted wounds.

Web delivery is a deceptively different model.

Re:What is a website?

[AJWM]

Don't kid yourself.

Management views the website, with its browser interface, as little more than a modern face on the old IBM mainframe application with its 3270-terminal interface.

The 3270 was a page-mode text terminal that let the app define fields the user filled in or checked off and then click a 'submit' key to send the form's contents back to the mainframe. Sound familiar?

That the back end now may consist of apps or components spread across several boxes, vs different apps and components in the same big box, isn't that huge a difference. (Just because the apps may have run on the same mainframe doesn't mean the whole thing was monolithic from an application sense.)

Re:What is a website?

[ekidder]

Funnily, it sounds very familiar. Perhaps that's because one of my jobs is to convert TN3270 applications (written in Taskmate) to use apache and mod_perl.

It ain't pretty.

And the surprise was?

[sbrown123]

".NET is finding a sweet spot for programmed user interfaces, while J2EE continues to enjoy its sweet spot for server-side applications."

So its just business as usual?

BEA is a joke

[signe]

Their software is way overpriced. And it's not like they provide support to make up for it. Every time I've had to deal with their support people, they're incredibly braindead.

I'm actually starting to evaluate Eclipse as a replacement for Weblogic Server (BEA's software) across our organization. Should save something like $17k per installation. Anyone have any experience using Eclipse?

-Todd

Re:BEA is a joke

[teetam]

Isn't eclipse an IDE that is competing with Forte?

As far as I know, JBoss and Jonas are the free, open-source J2EE appservers competing with BEA. May be you meant one of them.

Re:BEA is a joke

[signe]

Possibly. Like I said, I've only started evaluating options. I was pointed at Eclipse, and a quick look at it looked like it might be both an app server and an IDE.

Thanks for the pointers however. More work, more fun. More acclaim when I save us hundreds of thousands of dollars.

-Todd

Re:BEA is a joke

[Anonymous Coward]

I don't know which eclipse you're talking about, but if it's the same one at eclipse.org, the site states teh following:

The Eclipse Project is an open source software development project dedicated to providing a robust, full-featured, commercial-quality, industry platform for the development of highly integrated tools. It is composed of three subprojects, Platform, JDT - Java development tools, and PDE - Plug-in development environment. The success of the Eclipse Platform depends on how well it enables a wide range of tool builders to build best of breed integrated tools. But the real vision of eclipse as an industry platform is only realized if these tools from different tool builders can be combined together by users to suit their unique requirements, in ways that the tool builders never even imagined. The mission of the Eclipse Project is to adapt and evolve the Eclipse Platform and associated tools to meet the needs of the tool building community and its users, so that the vision of eclipse as an industry platform is realized

Eclipse is not an app server. It's a development environment with lots of pluggable features. I'm no expert on eclipse, but eclipse uses a different model for GUI development. Rather than use swing, which prefers to have a consistent cross platform look, eclipse wrote their own native libraries and wrapped java around it. This way you get very nice GUI with more low level control. for example, things like fonts and panes are handled differently and require memory management. If you create a particular gui widget, you have to free it up once your done.

Re:BEA is a joke

[RevAaron]

I'm afraid I don't follow you. Eclipse is an IDE. Like JBuilder or NetBeans, not an application server. Sure, you could build web applications from scratch or using a more primitive product using Eclipse or any other IDE or editor, but Eclipse doesn't replace Weblogic Server any more than emacs.

Maybe the support people don't follow you either, because you don't have a handle on what you're talking about? Maybe it seems overpriced because you're buying a web app server to do the job of an editor? Perhaps I'm being too harsh, but just some ideas.

Re:BEA is a joke

[obsidian head]

Eclipse is an IDE framework. Look at JBoss instead, for what you want.

Re:BEA is a joke

[obsidian head]

Interesting. Just Google for 'j2ee server' and hit the "I'm feeling lucky" button. It'll take you straight to JBoss.org.

Re:BEA is a joke

[(H)elix1]

I see others have pointed out the IDE (Eclipse) != App Server. The integration can confuse folks, however. IBM is spending big bucks to make Webshphere Studio -- Eclipse with WAS debugger and build integration their main development tool.

Anyhow, I've found that IBM will cost you just as much.... except you get something that is much rougher around the edges. Want EJB 2 support? IBM should be out with their first cut in another month or so. BEA went gold within a month of the spec coming out last November. My benchmarks showed it is not even close to a fair fight with the current v4 or beta5 vs BEA.

Solid EJB containers cost real money. Oracle is getting cheaper -- just starting to work with it. Sun looks like they are giving one away for free, but not sure how solid that one is. JBoss gets pretty high praises from those in developement.

Low cost IDE's for doing EJB work are the bane of Enterprise Java Bean developemnt. You end up spending big bucks 2K+ for an IDE like JBuilder, or you get something that works for only one App server. If JBoss expands the eclipse tool the same way IBM did, it will be a nice kit.

Re:BEA is a joke

[MCRocker]

If you want something that actually is a development tool with a built-in Servlet Container, then check out Simplicity Enterprise. It uses a drag-and-drop metaphor that is much easier than Eclipse. Since it has a built-in Jetty server (the same one used by JBoss) with Execution-On-The-Fly(TM) technology you can test as you develop without having to re-compile! Definitely checck out the video

Disclaimer: I used to work for Data Representations, Inc., so I'm a little biased :)

So it will be, write twice...

[aero6dof]

write twice, debug everywhere * (new releases +license update + security gaffes) * ms_reboot_multiplier

Uses

[Tablizer]

(begin list of real-world uses for web services)









(end list)

Re:Uses

[demian031]

i don't know what it is about the ppl that post here. maybe you've got your own little world mastered (be it perl, C kernel hacking or whatever) but you think you are experts on everything. ...kinda weird.

anyways, i'm doing some web-services stuff at work using M$ MapPoint mapping (web-) services (http://mappoint.net). there are a few kinks to work out but it seems to be a good fit. we're pretty much a java, j2ee (webLogic) shop.

if you've done any geo-coding it's a pain to get all the data on cds, load it into some database, slap some sort of server on top of it (3rd party or in house), etc... it's a lot easier to just call a web service.

Re:Uses

[Tablizer]

(* anyways, i'm doing some web-services stuff at work using M$ MapPoint mapping *)

I have not seen anything there that could not be done with HTTP get or post parameters (zipcode=12345&storename=buzzwords_R_us, etc.) and returning XML, comma-delimited lists, or a map image URL.

Here is a odd example on that site:

"As an example, let's say you are a waste removal company and you have uploaded the locations of all your dumpsters to the MapPoint .NET service. Along with each record, you have uploaded the pick-up schedule for each dumpster. You can now perform a proximity query to find out which dumpsters in your northeast territory need to be emptied next Wednesday. MapPoint .NET makes it easy to request a map of these returned POI."

Sounds like a bit more than a "map" service to me. It is MS trying to host *other* biz functions also (such as scheduling) it sounds like. Just another way for MS to get into your pants........pocket.

Re:Uses

[glwtta]

oh come on - there are great real world uses for web services. for one, UDDI, WSDL and SOAP are three more acronyms that make any resume look cooler!

Re:Uses

[Rombuu]

There are lots of great real world uses for web services. Do any sort of integration work. Hey, I have a web front end for ordering that needs to talk to my ERP instance on the back end to check inventory before I confirm an order. Do I

a) Code something in some language that does the lookup usually via some proprietary adapter or
b) Use UDDI to lookup the WSDL. Create SOAP call to do the lookup. Now if I ever change my web front end, hey, I've got something that works no matter what web server I use as long as is it can do SOAP calls.

All this web services stuff is standard based and reduces vendor lock in. I don't know what all the resistance is to it. Anything that easily lets Apache talk to SAP talk to Siebel talk to by RDMS talk to my logistics system is OK with me.

Re:Uses

[mewsenews]

STOP!! the irony!! too much!!!! ggaaazgzzzzz....

Re:Uses

[blackcat]

It's funny. Laugh.

Re:Uses

[Anonymous Coward]

I have not seen anything there that could not be done with HTTP get or post parameters (zipcode=12345&storename=buzzwords_R_us, etc.) and returning XML, comma-delimited lists, or a map image URL.

Well...no. But, then again, I've never seen an office suite that couldn't be hand-coded in assembler.

Re:Uses

[samael]

Let's see, I frequently use web based calendars, email, discussion forums, etc.

I use Livejournal, which has both a web based front end and a Web-API based one (which I prefer posting from).

I've communicated from a Windows based middle-tier to a tomcat web server using SOAP.

There's an awful lot of places that having web services is/will be handy.

Re:Uses

[TheAncientHacker]

And nobody could ever have a use for a computer in their home. Just ask the founder of DEC.

Those who find uses will get rich on them. Those that don't will whine about why the world changed out from under them.

Re:Uses

[GrayArea]

Well, I built a product configurator service (think Dell's build-to-order configuration of computers) for a (very) large computer manufacturer that exposed its functionality as a web service. There were at least five applications downstream including a couple of web front-ends in Windows/ASP that used these services and were all written by other teams in other divisions. There would have been no easier way for us to integrate this many applications on a common configuration engine without web services.

I'm convinced that web services is a better way of doing things when you need to integrate coarse-grained components on different platforms.

Re:Uses

[iabervon]

Web services are a great idea for intra-company stuff. If you've got some application that everybody needs to use on the same data that doesn't work well in text (say, a project management program), you could either write a GUI for each platform and a protocol for connecting to the server for shared state, or you could have an internal web server running the service, skip both of these steps, and just code behavior.

For a lot of things, I agree that web services don't make sense on the open internet; there are a whole lot of things that ought to be static pages that get done with jsps (see shop.usps.com, for example).

I think the real test for whether something should be a web service is whether the users all want to interact with each other. If so, it makes a lot of sense to have a web service. If people are just doing their own thing and don't want to interact, web services don't make a lot of sense.

eat your cake and have it too

[demian031]

word up, python rules! i wish everyone else would see the light!

...python for the jvm
http://www.jython.org/

...python for the clr
http://starship.python.net/crew/mhammond/dotnet/in dex.html

BEA's ploy to keep MS off their back

[Infonaut]

Unless I missed something huge here, the comment: ".NET is finding a sweet spot for programmed user interfaces, while J2EE continues to enjoy its sweet spot for server-side applications" in no way represents BEA giving ground to Microsoft.

BEA makes money by selling WebLogic, which is an environment for developing server-side apps. He seems to really be saying, "Look, Microsoft, you go mess around with programmed user interfaces, because server-side development is our turf."

Openly defying Microsoft doesn't usually work, but clever companies have managed to stay out of Microsoft's sights by promising to play well with Microsoft and stay in their niche.

Re:BEA's ploy to keep MS off their back

[alienmole]

Dietzen's comment is aimed at his customers and investors, not at Microsoft. He's simply saying "I hereby concede everything that doesn't matter to us", in a rather pathetic and transparent attempt to stake out a space within which BEA can still matter.

The fact that he's coming out and making such a statement so early in the .NET lifecycle is presumably an indication that BEA is under big pressure from customers and/or investors to come up with a coherent story of where BEA will fit into the brave new .NET.

I'm not sure that laying your head down dog-like on the ground and exposing your jugular to Microsoft is really the way to stay in business, though. BEA may think Microsoft doesn't have a sweet spot on the server, but that's a bit like the old joke about dinosaurs: it's OK, dinosaurs won't eat you, they're herbivores. Yeah, but do the dinosaurs know that? I don't think Microsoft knows it's not supposed to have a sweet spot on servers, and Dietzen's position isn't going to make a whit of difference to Microsoft.

Re:BEA's ploy to keep MS off their back

[The B]

I'm betting BEA willbuild in .NET capability in Weblogic. Watch and see.

\dotted

[ramzak2k]

slash dotted ! I did it ! the very first one to kill the sys-con beast "Initiating server query ... Looking up IP address for domain: www.sys-con.com The IP address for the domain is: 207.178.67.98 Connecting to the server on standard HTTP port: 80 The port is closed, so our connection attempt was refused. Query complete."

Fact and fiction

[f00zbll]

Can't get to the damn article, so rather than talk about what Scott Deitzen said, I'll talk about what I do know. There's a lot of confusion about web services and most of it is marketing BS. Web services the technology, language, and framework isn't BS. Loosely, it an attempt to move away from tightly integrated systems. What does that mean? To me that means rather than have clients and servers all designed specifically to use a limited set of protocols, applications are designed to see each other as services. where a traditional client server application like point of service connects several dumb terminals to a central server using a proprietary protocol, web services defines standard ways for discovering and using services. A service is defined by WSDL (web services description language) which defines the server, port and message structure. when two applications need to talk to each other, they look at each other's WSDL to figure out how to transform a message into a desired binary format.

IBM has been steadily putting lots of good code and ideas about how to approach web service. Having read through most of the white papers on microsoft's site and worked a little on .NET application which uses SOAP and WSDL, microsoft's model is not ready for prime time. IBM recently release a Web Services Innvocation framework and web services toolkit.

there are also other efforts like castor, which are designed with web services model. IBM has a protocol called Web Services Flow Language, which borrows ideas and techniques from pi calculus. Web services as a technology will mature, but not at the rate microsoft claims or in the way microsoft thinks.

Re:Fact and fiction

[f00zbll]

Here is an article that is more balanced and realistic about web services.

Still lacking specifics

[Tablizer]

(* The modding up of such stupidity is going to finally drive me away from Slashdot, I think. It's hard to blame the morons who post gibberish at nearly every opportunity, when, with useful moderation.... *)

I notice in all your counter-rantings that you never described an *actual* application, and how .NET or JavaBloatX do it *better* than HTTP, SOAP, FTP, ODBC, etc.

There are plenty of existing communication approaches out there. Sun and .MS are simply trying to out-acronym the other in the hearts and minds of PHB's.

In the mean-time, we have open existing K.I.S.S. protocols to get real stuff done without selling out to the battling fat cats and getting shot up in their cross fire.

Let them battle by themselves, we don't need 'em.

Re:What about PHP?

[Real_Mce]

Yes I would have to agree PHP is truly amazing and with PHP-GTK you can actually write cross platfom applications.

from the Department of Redundancy Department

[Tablizer]

(* You, however, said "web services", and since you're tossing SOAP into it now as if this unrelated to "web services", coupled with your pseudo-rant, I'll assume you're not actually talking about "web services", and are talking about specific frameworks for implementing them. *)

I suppose a definition of "web services" is needed to clearify this discussion.

If web services is "using the web to communicate between 2 or more different machines", then "web services" is actually *redundant*, because that is precisely what the WWW is for, isn't it?

It would be like calling something the "data transfer network" or "transportation vehicle" or "fly-capable airpline" or "floating boat" or "kill-capable military"[1], etc.

[1] Well, I suppose there are some really lame militaries out there.

Re:from the Department of Redundancy Department

[Tablizer]

(* What it can do is produce a stable, well-architected, extensible application where you previously had 40,000 lines of Perl that burnt your eyeballs to look at. *)

So it is replaced by 50,000 lines of bloated Java and .NET code.

I agree that some standardization is needed, however that can be done *without* marrying Bill Sun Gates. You really don't have to bend over to cough, you know.

Pretty much all you need is an XML parser and an HTTP character encoder/decoder (for ampersands, etc.).

Re:from the Department of Redundancy Department

[ednopantz]

So it is replaced by 50,000 lines of bloated Java and .NET code.

Have you actually done anything with .NET? Nasty, unreliable screen scraping is replaced with about four lines of code.
The .NET framework is huge, but so are modern hard drives.
The application programmer has an easy time of it:

1) declare the web service object
2) instantiate it
3) call the web method, returning something you want
4) dispose of the web service object
now do something cool with what you got back.

Real world uses: how about an order tracking system for companies with hundreds of locations that don't want to lease T1 lines to tie everything together.

You can return data with web services, or use it just like a piece of middleware to write to a database.

Get a DSL line at each office, set up an IIS server, throw SSL on it to encrypt your soap communications, use client certs to identify your client workstations, firewall and IDS like hell, and you have a fully functional distributed order processing system for $100 a month per location in connectivity and $2000 for a decent managed server. No kludgy DHTML user interfaces. Reliable client side validation, a fancy Windows form UI, no expensive VPNs, no T1s, a cheap, fully functional distributed application.

Yeah, "you have to buy Windows... bitch, bitch, bitch..." yeah, but since Windows is 90 some odd percent of the OS market, and probably 98% of business market, who cares? You have to get tied to MS, but businesses get tied to vendors they resent all the time.

For people who need real world applications, want them done good enough and cheap enough and aren't engaged in the Linux Jihad, .NET is a killer platform.

"This guy must be an MS drone," No. A consultant who uses .NET to offer my clients services that the HTML/Perl solutions can't compete with.

Re:from the Department of Redundancy Department

[Tablizer]

(* Nasty, unreliable screen scraping is replaced with about four lines of code. *)

Nice little false dichotomy: Screen scraping versus .Net.

(* Real world uses: how about an order tracking system for companies with hundreds of locations that don't want to lease T1 lines to tie everything together. *)

Have central servers. You don't need distributed fat clients for that. For those operations that don't fit a web-based interface, then use something like XWT or SCGUI. If that is not feasible, then use VB-like client interfaces for those parts, but avoid bound controls. It can even be self-updating if you do it right.

What's really gonna happen is...

[3seas]

Cornering the Autocoding Market .net is just the setup.

Some info on Autocoding

Re:What's really gonna happen is...

[Zurk]

dude. next time post under your own logged in name when posting such a godamn incoherent post.
autocoding aint going anywhere. not anytime soon anyway. and it has nothing to do with .NYET

You think BEA is that weak?

[Infonaut]

... a rather pathetic and transparent attempt to stake out a space within which BEA can still matter.

Do you mean to imply that BEA is scared with good reason, or that they're scared for no good reason? I guess what I really want to know is whether you feel BEA is weak, or J2EE is weak, or both?

I'm not sure that BEA is exposing its jugular, but I definitely agree that playing the "we can all get along" game with MS usually leads to catastrophic meltdown. Only a few software companies can truly say that they've been able to successfully partner with MS without being chomped on hard. But using the McNeally approach isn't always successful either.

It seems that one of the most vexing business challenges of this age is "how do we compete in the same market as Microsoft?"

Re:You think BEA is that weak?

[alienmole]

I don't really know how good BEA's reason to be scared is, I was really just saying that Dietzen's statement seems to demonstrate that they're scared, or perhaps in denial.

The problem with Dietzen's statement is that it's not really plausible - he's saying that BEA's OK because Microsoft won't find a sweet spot in the application server space. The obvious problem with this is that Microsoft will certainly find a sweet spot in that space, it's just a matter of time. So Dietzen's statement, instead of showing why BEA has a future, actually calls its future into question. He's drawing attention to their weakness, which seems like a questionable strategy. Of course, the weakness is fairly obvious, but if that's the best defense he can come up with, then man, BEA's toast!!

I'm not saying it's inevitable that BEA will be trampled by Microsoft, but a comment like Dietzen's seems pretty revealing as to what he thinks of their prospects.

It seems that one of the most vexing business challenges of this age is "how do we compete in the same market as Microsoft?"

I agree - as a software developer, I've run into this issue personally. It's a big reason I focus heavily on free software, where appropriate, and recommend it to my clients, many of whom have been burned by Microsoft's shifts in direction, lack of standards, and lack of openness. In a sense, Microsoft has been one the best motivating factors for free software. I doubt, for example, that IBM would be doing so much free software if it weren't for Microsoft.

In a way, BEA is actually caught between Microsoft and free software. Open source J2EE servers are pretty much in their infancy, but the problem for BEA is that the "intellectual property" inherent in today's application servers is really not that difficult to duplicate. So you get free servers like JBoss, and cheap servers (with source) like Resin. I doubt many BEA clients are quite ready to jump ship today, but over time, the standardization and network effects of free & open solutions can become quite compelling. I don't think time is on BEA's side...

MS Model not ready for primetime?

[Otis_INF]

LOL :)

And why's that? WSDL isn't MS proprietry language, it's a standard defined by many companies, and should be used with UDDI, currently in v2.0. .NET is a framework which IS ready for primetime, because it offers all the functionality needed to build mission critical applications which have to serve thousands of users.

So, f00zbll, show me the beef where .NET falls short when it comes to delivering what's promised.

ps: I develop a lot of .NET software, I know what I'm talking about.

Re:MS Model not ready for primetime?

[f00zbll]

I don't claim to every single thing about how microsoft sees web services, but I've tried to playing microsoft's MapPoint.NET and went through most of their online docs.

If you go on MSDN and look at the MapPoint.NET docs, examples and the problems people have been having you'll concede there's mixed message coming from microsoft. What do I mean? First off, the way MapPoint.NET SOAP works, it requires authentication using NTLM. Some will recognize it as IIS Challenge/response. Problem is when you try to connect to MapPoint.NET soap servers, it may take multiple connections. In MS MapPoint SOAP press release they claim to be better than Mapquest. I've used both. With Mapquest it takes one connection because the authentication and request happen with the same connection. Not only that, you can get a remote URL to a map. With MapPoint.NET you get map encoded in base64. This means any developer integrating MapPoint SOAP has save the image locally and manage those resources. If MS went to a stub model where the request generates a stub to an image, the response would be faster and reduce the traffic to the client's server. The application can print the URL to the browser and the browser would get the image from mappoint servers. You might think the MS model is better, but having worked with several mapping products, the MS model of a "web service" sucks. Following the MapPoint SOAP model would mean more work and more hardware to support what would be a simple application with Mapquest. Now I don't that this is how MS sees web services, but MapPoint.NET soap is the first official .NET service from MS that I know of.

Now lets look at this closer. MS recommends using UTF16 encoding for SOAP messages, but digest authentication with IIS requires base64 encoding. Having the ability to use different encoding provides great flexibility, but doing that over HTTP seems lame to me. Sure SOAP is not dependent on HTTP, but according to MapPoint SOAP in its current form it is tied HTTP. MS could have easily used encryption to hash request and make req/resp one single connection, but instead they chose to turn on challenge/response.

There were several questions on msdn about using other languages with MapPoint SOAP. If someone is trying to integrate mappoint soap with something written in C or C++ on unix, aix, irx, or vms it would be a pain. when developers asked about other platforms and languages, the response was "it's not supported, you're on your own." How's mappoint soap suppose to make a developer's life simplier? If MS is really interested in making loosely coupled web services, then it should support other languages and platforms, so that someone on a vms mainframe who wants to add mapping to an application can do it without writing their own SOAP, UDDI and WSDL drivers. Technologies like castor and IBM's WSIF appear more balanced and flexible. Even though MS uses standard protocols, their implementations and use is different.

Lets not confuse implementation with protocol specification please.

this article on ibm is honest about web services and it's short comings.

Re:MS Model not ready for primetime?

[LibertineR]

"I don't claim to every single thing about how microsoft sees web services"

You should have stopped right there. Save your PrimeTime predictions for things you know about.

Re:MS Model not ready for primetime?

[aeoo]

"I don't claim to every single thing about how microsoft sees web services"

You should have stopped right there. Save your PrimeTime predictions for things you know about.

You don't need to "[know] every single thing about [something]" to talk about "things you know about".

If you think that knowing every single thing about something is even possible...never mind. f00zbll was being humble. Do you understand what "humble" means and why it is good to be humble?

Re:MS Model not ready for primetime?

[f00zbll]

If you have experience and facts to back up your assertions, I'm more than willing to admit my own ignorance/misunderstanding. Saying something without one piece of technical information is pretty easy. All of the facts I stated are verifiable on mappoint soap message board on msdn. I'm not the only person who finds mappoint.NET soap to be frustrating for non-windows platform.

I've also used the non-soap version of mappoint and in my mind it's a step backward feature wise. With the older version, you can get a URL to a map, whereas the newer version requires the application to send it directly to the browser, or save the image locally. This causes more problems and work for the developer because it means there has to be a thread to clean up the images, or set an arbitrary number of images and over write the old ones. Other mapping applications support the ability to get a remote URL like the older version. As far as I know, MapPoint.NET SOAP is the version .NET service released by MS. Of course no one knows if that is going to be the dominant model for MS, but from this particular model, it's far from enterprise class service. I a person wanted to integrate mapping into a site that gets 500K map hits a day, retrieving the map every time is going to be expensive bandwidth wise and require a lot more work to build a scalable website. Others may have different experiences, but that's what I've seen first hand.

I don't proof read cuz I'm lazy

j2ee and .net stuff

[fedor]

I still don't understand the real advantages of this kind of stuff. Honestly I've tried to understand the whole j2ee stuff, and I still can't. I haven't tried .NET and I will never try it, because all the books about it with titles containing words like 'Enterprise', 'Solution', 'Deploying' and 'Integrated' just make me sick.

Big companies just try to sell something 'good' for software development. The problem is, that they sell it to suits who just believe the marketing crap of sales people. Most of the time there are no technical people involved in these decisions and people like us just have to work with it, because 'it's good', 'scalable', 'integrated', blkahblahblah...

I really hope all this crap is over in a couple of years, and we can just use whatever we want to use and know is good. Getting experienced in using progamming languages and tools is the only way to see if things work. My experience is that in projects where they used stuff like J2EE, SOAP and COM, a lot of people where involved (a lot of suits) and the projects last long....over 2 years for a simple web based application!!! Arrgrhgh!! I've seen a lot of those projects and nobody overthere knew how to query a database using SQL how to use TCP-sockets, or what's the HTTP-protocol all about...They laughed at me because I did not want to use Visual Age for Java and I used javac and vi instead. "In Visual Age I cannot not see on which line parse errors occur" was my answer, when they asked what was better about just using the command line. I had to explain the word 'parser'.....

People on these kind of big projects just had a one week course and some experience with Excel macro's. And who sells those courses....???


I think, it's all corrupt...

fedor

Re:j2ee and .net stuff

[floydman]

i totally agree with u, infact i face the same problem, lots of coders nowadays have the attitude of learning some API's (in a 2 week course maybe), and lack major concepts (like what is a parser in ur case), and these kinds of coders just push an unreliable technology forward, leaving all the good stuff to rotten...
i agree

Re:j2ee and .net stuff

[BigBadBri]

I agree about the suits, but the flecibility that using an appserver (I'm using Tomcat for development) to allow me to write servlets that talk over TCP to a custom java application that does all my DB queries, with tools that will allow me to create my application in around a month - all this makes me thank Sun for J2EE. I'll never program in D-flat, though - why bother to learn another bloody language?

Re:j2ee and .net stuff

[fedor]

Tomcat indeed works fine....it just does the same as the big and expensive commercial bloatware. But do you really use EJBs and other j2ee related stuff? My experience is that software which uses this is (most of the time) slow. You really have to know what you're doing when using existing frameworks like this.

Regarding learning programming languages, I don't bother learning something new, but when I already know how things have to work and know how to implement it, why should I use a framework which hides the internals?

fedor.

Re:j2ee and .net stuff

[Tablizer]

The problem is, that they sell it to suits who just believe the marketing crap of sales people......I've seen a lot of those projects and nobody overthere knew how to query a database using SQL how to use TCP-sockets, or what's the HTTP-protocol all about...They laughed at me because I did not want to use Visual Age for Java and I used javac and vi instead.

Agreed. It is mostly snake-oil. Everybody is looking for the silver bullet to avoid having to learn things like SQL and HTTP. The problem is that in the end you end up having to fight the "magic" IDE and goofy API's to get what you *really* need the hard way.

It is all just a big glass of whine: Delays the innevitable, but makes the innevitable worse when it comes.

The 80/20 [1] rule is a universal fact of life in IT. Perhaps someday somebody will discover the magic IDE/tool/API that gives both, but until then one should learn to avoid the BS.

[1] The 80/20 rule is that tools that speed up the first 80 percent of the project tend to slow the last 20 percent to a crawl because they don't provide the granularity of control that you need.

Re:j2ee and .net stuff

[j3110]

Agreed. You don't have to use all of EJB's features that slow down the program on simple programs to get some benefit. I like the security model. Sometimes you need row level security on data, and EJB's make that easy (a couple of lines of code). I don't like data in inconsistant states at all, so transactions that work across multiple database servers is great and it's all configurable in an XML file so I can remove transaction points without toying with the code. DBMS independance is my favorite selling point. If I find a faster database server than MySQL (with InnoDB for transactions) I could change it with one file. If you want to "Upgrade" to Larry Elisson land, or you have to write it to be compatible with a system that is tied to MS's not quite SQL server, you won't have to change your code when the legacy system is reimplemented into the new J2EE system.

Honestly, some people need to rethink how much performance is really worth? If you can save time by not having to write transactions, security, and testing both (which is a REAL PAIN), you can spend the money or time you saved there to either optimize or buy more hardware (since it supports clustering without you having to write any code, just more machines).

.Net ? Ha !

[Aelentel]

Interoperability ?
okay, i'll take on that word.
Interoperability does not restrict to web services,
just questions : CICS ? 3270 ? CFT ? SNA ?, MQSeries ?, etc... etc...
All of these are a nessecity in banking and some other business webservices.
so when .net will support an API to these, and when it'll run on other platform than microsoft (IBM WhinterHawk for example)
maybe i'll consider .net as 'interoperable'.

ps : sorry for my bad english :)

Re:.Net ? Ha !

[spells]

MS does support most of those "standards" through Host Integration Server. Running on multiple platforms means it's portable, not interoperable. MS doesn't claim .NET is portable.

Better Framework

[jimmyCarter]

I hope I'm not starting a flame war, but I've written apps in each language and the framework for .NET is the the Java Class Library given time to mature and improved upon weak spots. Reflection kicks RTTI's ass among other minor, but effective improvements. I'm not trashing Java, I enjoy programming in each language, but C# is what I'm using more and more to build with.

Microsoft are playing catch-up

[JeremyMcGee]

What Microsoft have done is to concede that virtual machine architectures are the only way forward for enterprise development on Windows. They tried ActiveX/COM/COM+ as a way to do distributed development. That failed.

Microsoft won't back down from .NET any time soon, so comparing the architectures is absolutely valid.

It took Sun years to get Java to the stage where a VM is considered to be a good way of deploying programs. .NET has at least got them to move forward with technologies like JAX-RPC.

Web Services When I Can't Even Get My Email?

[Master of Transhuman]

How does Microsoft (or anybody else) expect major corporations to run mission-critical stuff over lines provided by ISPs who routinely oversubscribe their boxes and undersubscribe their bandwidth?

I can't reliably get my Usenet newsfeed without "Connection unexpectedly closed by server" messages.

Anybody think you can run General Motors or any bank on that basis? Anybody think any ASP isn't just going to be an ISP with a new acronym?

In a way, until the phone companies get us that infinite bandwidth they were promising a couple years ago, this is good because it will probably kill Microsoft when it becomes apparent that none of this will work for reasons entirely outside the issue of which programming language or object broker is used.

cross platform ui

[primus_sucks]

.NET is finding a sweet spot for programmed user interfaces, while J2EE continues to enjoy its sweet spot for server-side applications.

That's great if you're only targeting MS platforms. If you really need high performance, mostly crossform UI, the eclispe project's SWT library seems like a good choice. If you need totally cross-platform UI, and you're willing to sacrifice some performance, just use swing. In any event, architect your software such that your logic and server communication is separate from your UI, so you could replace the UI with the least amout of effort.

Re:cross platform ui

[nicestepauthor]

I've been using eclipse for awhile and I like how SWT works in that application. You'd never know that app was written in Java if nobody told you. However, it does need more memory to run than Visual Age for Linux did. I'm well aware of this because I have developed on both on a Pentium with 64 Meg of memory (an Aptiva, so you can't add more memory). So the UI seems to be only part of the problem with Java.

I have used Swing apps like JEdit on machines with a lot of memory and have no complaints. They look as good as a Windows application and performance while they are running is decent. Startup time is slow on any Java program though.

I have also made my own lightweight components which are much less elaborate than Swing but seem to me to work well enough on a Pentium. Making your own Java ui components is easy and even fun, and I've wondered why more people don't try it.

Economcis of Webservices in a declining Economy

[linuxislandsucks]

Hmm CTOs making blathering comments..so when is he going to get fired?

But seriously, the third choice is either .NET or J2EE using oipen source....With the world ecomony declining for the next 4 years in resposne to USA's declining economy to election year ...Worl midlevle companies which encompass 90% of world GNP do not have themony to deploy boht webservices platforms from high priced vendors..

Thus you will see a marked increase in opensource deployment of boht webservices platforms rather than vendor offerings..

Open Source is entering the New Economic Revolution! Are you ready to Rock?!

.Net ads on slashdot?

[tongue]

Does anyone else find it strange to see a microsoft ad on slashdot?

.Net vs. J2EE vs. PHP etc

[theolein]

As someone who doesn't have much of an idea as to why webservices are needed (I've read the articles and looked at some examples) because I often just think that the whole idea of web services is like a minutarised packaged version of the ASP boom and flop a few years ago and that there are definitely other ways to do this. To me it seems as if the programming side is as much a marketing push by IBM,MS and others to generate more business than a real innovation.

To me the idea of webservices seems to be the ability for a client of some kind (could also be a server type of client) to send a stateful request for info to a server of some kind. Since developers have been working around the Browser's lack of state for years and non browser applications can use a myriad of protocols etc to communicate to their server, such as XML-RPC,RMI,CORBA or whatever.

Turning SOAP and UDDI into a universal standard is nice but can anybody tell me why companies trying to sell web services are any different from companies that tried to make a living (and flopped) with the ASP thing a few years back?

Re:.Net vs. J2EE vs. PHP etc

[f00zbll]

send a stateful request for info to a server of some kind. Since developers have been working around the Browser's lack of state for years and non browser applications can use a myriad of protocols etc to communicate to their server, such as XML-RPC,RMI,CORBA or whatever.

No where in the current WSDL, or SOAP specification mention stateful management or framework. If you look at IBM's perspective on web services, it includes both stateless and stateful applications. If you look at MS's application center and application server, it explicitly says:

• Accessing Stored Data Unlike ADO, which makes it easy to build heavy-touch clients that don't scale well, ADO.NET is biased toward building light-touch clients. An ADO.NET client uses forward-only, read-only cursors to read data. Stateful server-side cursors aren't supported, so the programming model encourages short connection lifetimes. Clients that read and process data directly can use the ADO.NET DataReader object, which provides no caching for returned data. Alternatively, data can be read into a DataSet object, which acts as a cache for data returned from SQL queries and other sources. Unlike an ADO Recordset, however, a DataSet cannot explicitly maintain an open connection to a database. the original page

I still a lot of conflicting/confusing/contrasting marketing bs about what web services really means. Until everyone agrees and makes all the necessary protocols and frameworks free, web services is going to have a hard time finding wide acceptance.

Re:.Net vs. J2EE vs. PHP etc

[MarkWatson]

Good points!

I have used SOAP and UDDI a fair amount and they are good technologies for building loosely coupled systems with existing services, but in today's economy, I question how widely used they will be.

Some companies are experimenting with SOAP based services; Google is a great example of this (as a developer, you can get a free use license for a 1000 queries a day for search, spelling correction, etc.)

I have been playing around with the idea of providing NLP web services now that I have a dedicated server to use, but I am more than a little sceptical that individuals or companies would pay, for example, 1 cent per transaction.

-Mark

Re:Windows has a place

[zapfie]

Here, this comic may help...

Google (N/T)

[jpmorgan]

See subject.

show me the bennies

[Tablizer]

(* Well...no. But, then again, I've never seen an office suite that couldn't be hand-coded in assembler. *)

I asked about *benefits* already. Show how the MS/SUN approach is also faster, better, cheaper, gets your shirts whiter, etc.

Drag and drop?

[Tablizer]

That is the point. C#, ASP.NET and VS.NET are the perfect combination if you want to make HTML based UIs. You just drag a button onto a webpage, double click and write the event handling code in C#.

That is kind of suspect since real HTML is flow-based and not coordinate-based. Thus, MS is probably breaking a standard or making up a new one, unless I am missing something here. (I suppose you could use style-sheet absolute positioning, but that was buggy on some browsers when I tried.)

(BTW, Remote via-HTTP GUI's are highly possible IMO. Web forms for biz apps don't work very well for non-trivial stuff IMO. There are technology drafts like XWT and SCGUI for remote GUI's. They just need a big-name push.)

Question for PyroPunk

[jgeelan]

as a Java developer who's migrated to .NET, what do you make of what Sun's Rima Patel has to say about Java and .NET? http://java.sun.com/features/2002/07/rimapatel.htm l

Re:Just say .no

[supton]

PHP : pretty hackish programming. Great for what it is good at (inlining simple code in HTML). Bad for anything else (enterprise apps). If you want something simple (but not decafinated like PHP) as an alternative to Java and .NET, the Python and Perl communties have quite a few options that might be better suited than PHP (Python moreso than Perl, for pure OO apps).

correction (Re:More Python Please!)

[Tablizer]

(* There are multiple division candidates for any non-trivial app *)

That should be "multiple orthogonal division candidates".

Unfortunately, modularity is often a dead-end goal IMO. Somethings you can split up into a nice black box off to the side, some things you just can't. Rather than chase a dead-end dream of boxing everyhing, we need to think about creating "virtual modularity". This is where you can create the modularity you need for a *particular* need. Relational technology comes the closest to this IMO.

OO just becomes exponentially complex when it tries these kinds of things because it is pretty much a one-dimensional technology.

OO is wonderful for those items that fit and change along that one dimension/border, but barfs on those that go against its main grain.

Re:Just say .no

[simeonbeta2]

Please don't knock php as "hackish" programming and then praise perl. I could note that some of the most obfuscated code i ever see is written in perl. I won't. :)

Instead i'll note that the quality of the code in any programming language is all in how you write it. It's true that php lends itself very well to dropping snippets of code into html. As a professional web developer, however, i use php in "enterprise" applications everyday. phplib provides me with a templating class so that content and programming are completely separated, a OO html widgets class so that form elements are generated and automatically validated with client side javascript, a database abstraction class (switching from a prototype in mysql to the application with MS-SQL db is no problem), and fine tuned session and authentication classes.

Even php as a language itself works well in an enterprise context: c like syntax means training new new graduates is easy, OO (as of version 4, at any rate) is powerful, and the built in libraries are comparable, if not superior, to other scripting languages (regex and string/array/stack/list manipulation, everybody's got, but flash generation, pdf interaction, SOAP, xml-rpc... with php the riches go on and on).

Finally, I find mod_php to be very comparable in speed of execution to mod_perl (benchmarks here)with superior readability. All this to say: opinions are fine, but back them up with real world experience and real data. If you have had experience with php in an enterprise level application where the language or runtime failed to meet your needs, please share with /.! Otherwise, please don't run down my web-app language of choice with hearsay simply because you learned something else first...

Regards
Simeon

Re:RTTI is for C++, not Java, you cockfoster

[jimmyCarter]

You're right genius.. it may be in java.lang.reflect, but check out chapter 11 of Eckel's Thinking in Java: Run-tim type identification (RTTI). It's a concept..

I love the high-school slur when you use cockfoster in the title of your reply.