Slashdot Mirror

> Unfortunately I don't think it comes close to something
> like lucene. But that's another kettle of fish.

You might want to check out Solr. It's a search server front end for Lucene.

Disclaimer: I work for CNET, who open sourced the project.

The website is at http://incubator.apache.org/solr
The wiki is at http://wiki.apache.org/solr

> Unfortunately I don't think it comes close to something
> like lucene. But that's another kettle of fish.

You might want to check out Solr. It's a search server front end for Lucene.

Disclaimer: I work for CNET, who open sourced the project.

The website is at http://incubator.apache.org/solr
The wiki is at http://wiki.apache.org/solr

It' right here, duh.

Role
I work 60-70% of my time as a member of the core consulting team here and the rest of the time on "IT" administration and management around the local office. I should note though that I am a software engineer first and fore most, but it so happens that in small businesses one must wear many hats. Last year I was also heavily involved in accounting activities and managed a marketing program.

Scale
I only have 30 workstations and 27 servers (only 2 are publicly accessible and 8 are in a RCF) to worry about presently:

Culture
It should be noted that my users are technically very competent, which is a totally different can of worms to you (I assume from your comments), but there are plenty of issues to guard against with too competent a user as well!:) The issues are just different.

Environment
Server OS: RHEL 3
Workstation OS: Fedora Core 4 and 2 MacOS X (those damn graphic designers/marketing folk!:)
VPN Server OS: NetBSD 3 (runs on an Alpha box)

Software Tools
SCM: Subversion http://subversion.tigris.org/
Issue Tracking: Trac, which integrates nicely with Subversion http://edgewall.com/trac/
Internal Documentation (for future growth): Trac's built-in wiki http://edgewall.com/trac/
Web Server: Apache, mod_python, mod_ssl, mod_dav, and all that good stuff http://apache.org/
Knowledge Base: OpenCyc (but looking for something better that is still open source)
Intranet Framework: Python 2.4/TurboGears/Apache/mod_python http://python.org/ and http://turbogears.org/
Authentication: Fedora Directory Server (LDAP)
Updates: Yum, up2date
Server Monitoring: Nagios http://www.nagios.org/
[Internal] Remote Access: ssh and Gnome/VNC for the rare visual task
[External] Remote Access (i.e. VPN): OpenVPN

Internal Tools
Fixed Asset Management: Rolled my own TurboGears Web/AJAX application that hooks into our accounting system (it took 3 days part-time).
Backups: Rolled own Python backup mechanisms including scripts
Deployment Tools: Using Python's autoinst http://autoinst.tigris.org/
Continous Integration: I have started using Bitten instead of using cron and shell scripts to launch Python distribution builds and tests on a nightly and "continuous" basis for immediate feedback - something I find invaluable.

Office Software
As mentioned in a previous posting using a good calendaring tool is a very good idea. My recommendation is the Calendar extension for the Mozilla suite of tools.

Either way, can you see anyone ever writing a library or API technology in JAVA, let alone an application that doesnt run like syrup on a cold day?

Yes. Batik, for example, is written in Java. Admittedly, it is much slower than Inkview in starting up and opening a file, or opening another file once started, but on the other hand, it supports filters (which inkview doesn't) and clips properly at image edge (inkview doesn't), and supports scripts and other niceties.

JAVA is over 10years old, and yet on P4 3.4ghz computers, you have to performance optimize even simple appliations to run reasonably well. Who in the heck thinks this is normal?

X86 assembly has been around since 1978, and even on a P4 3.5GHz computer you have to optimize to run well. I guess that means that assembly equals bad performance. It couldn't possibly be because "simple application" nowadays means something different than 10 years ago.

Not to say that Java is neccessarily a fast language, but saying that "you have to optimize to get good performance" is true for any programming language.

Apache does not fork a new process for each request in Windows, it actually just creates a new thread. (Source) Not that I would use it in Windows anyways though.

In fact, Apache on Windows does use threads, and not processes. From http://httpd.apache.org/docs/2.0/platform/windows. html:

Because Apache for Windows is multithreaded, it does not use a separate process for each request, as Apache does on Unix. Instead there are usually only two Apache processes running: a parent process, and a child which handles the requests. Within the child process each request is handled by a separate thread.

Well, there's Struts for one... ;)

Open Source Champion IBM is the single largest patenter in the WORLD

That's right. If you lack the cluefullness to observe the obvious in your own oxymoronic illustration: that IBM is no "Open Source Champion" but merely an opportunistic leach; you have clearly checked out of the reality department long ago. Software and business method patents do not have a "quality problem", they are fundamentally structurally unsound. Even Bill Gates understands this.

If you've spent any time at all with an IBM sales executive, you would realize that IBM's open source strategy is simply a way to bait people into using software which will hopefully segue into proprietary upgrades.

As RMS has so elequently and accurately stated over and over again: the greatest impediment to software development is not innovation - we have had plenty of that with no help from the USPTO. The greatest impedement to software development is the creation of large scale systems. Because of software patents, any software project of any substance must navigate a legal minefield. That is an impedement to progress, much more so than any threat to the pretensions of petulant greedy developers who think their little brain farts should feed their grandchildren.

FogBugz is great. Sure, it's a commercial system that you have to pay for, but it is easy to install, simple to use, has a very clean user interface and even has a philosophy. Believe it or not, the last point is the most important. The folks behind FogBugz seem to work really hard to adhere to the KISS pricipal and produce a superior product.

If you compare them to workhorses like Bugzilla, Fogbugz seems very minimalistic, but it turns out to actually be more useful that way. The guy behind the folks behind Fogbugz, Joel Spolsky, has lots of interesting things to say about the design of Fogbugz that are just good reading for ANY CS/IT person to even if you don't buy his product.

Another product that I tried out was scarab, which was appealing since it was a Java J2EE application from the same folks who brough us subversion. From a CS point of view, scarab is an interesting example of how to use turbine. Unfortunately, scarab is hard to install and configure.

Although the version of scarab that I tested was still a beta product that might not be quite so hard to use out of the box any more, it is interesting to compare it to FogBugz. Scarab had the kitchen sink approach that is so configurable that it could be set up to be every bit as complicated as Bugzilla or as simple as FogBugz. However this flexibility made it a nightmare to configure and administer. While you could, conceivably set it up like FogBugz, it would be hard to make it work exactly the same way and wouldn't provide the same ease of use... just the same limitations with an added level of complexity.

To summarize less is more... in quality and price this time ;)

It's not as if J2EE is the only way to do web application development with Java. Some times a Mack truck is needed, most of the time a pickup will do.

*Sigh* They are looking for Non Obvious Relationship Analysis (NORA) technology. It would be fairly easy to get up a OSS project to produce it. I actually tried to start up one (Ghandi/Custer) based on Globus GT4, but had to abandon it due to practical considerations involving not starving to death and having a roof over my head...

Most of the technology to do this already exists as Open Source projects. If I were starting it up now, I would probably try to combine an Open Source JBI /JSR ESB (ServiceMix http://incubator.apache.org/servicemix/) with GT4 and either Postgres or MySQL (Postgres has better spatial and language integration, and MySQL has that dual license issue, sophisticated features, and a large user base.).

ServiceMix has a built in rules engine. Properly reworked (an easy task with JBI service integration and drools architecture) it could easily cope with the demands involved (which are pretty extreme). The real trick is to have theoretical knowledge necessary to create the appropriate rules. A lot of OSS developers are working stiffs, and most non academics are not up to date enough on the Logic, Langua and Mathematics necessary to produce something like this.)

On the server side ...

• PDF::API2 for LAMP.
• Apache Formatting Objects Processor for J2EE
• I used PDFLib for ASP.NET but it now appears to no longer be open source. Curious as I didn't think that was possible. Goggling around found PDFSharp.
• PyX is for Python developers.

PDF belongs to adobe and to develop using it you have to pay them for their patents use. So if you want to distribute yourself some PDF that's OK but if you want to use any generating PDF or reading PDF programs you need to pay adobe the big money. And that's just leading to more and more lockin.

Utter rubbish. A number of different libraries capable of generating and working with PDF documents are available; for a free (as in beer and speech) Java one, look no further than Apache's own FOP.

Adobe's desktop applications (eg Distiller) are pay-for, yes, but there are no patents or other licensing issues; the PDF spec is freely available if you want to write your own implementation.

By keeping the text at a higher level of description as with HTML or an XML schema, the structure of the document is preserved. With PDF, paragraphs are broken down into lines of text and individual characters that are plopped down at specified coordinates. This breaks the document structure.

I would be more interested in a meta document format that also included suport for XSL-FO such as Apache FOP. This would provide a closer approximation of PDFs capabilities.

Maven has come quite a ways in the past year as well. If you're looking to ditch your overly complicated ant build scripts for organized simplicity with reports, take a gander:
http://maven.apache.org/
I'll go out on a limb and say it will be more important than eclipse in 2 years.

That is definitely going out on a limb

Maven has come quite a ways in the past year as well. If you're looking to ditch your overly complicated ant build scripts for organized simplicity with reports, take a gander:

http://maven.apache.org/

I'll go out on a limb and say it will be more important than eclipse in 2 years.

The Apache License is more of a BSD style license. There is no requirement to release modified source code.

A few other things from the article that need correcting: 1) not web 'scramblers' but web 'scammers' :-) and 2) The number 100 was a joke, I meant a number much larger than that but we don't talk about the number of machines that we release. For more info about our open source efforts and to see the code that we've released, see Code.Google.com

Chris

<?php
if (isset($_POST['name'])) {
    exec("echo 'your name is: '. $_POST['name']");
}
?>
<form action="./" method="post"><input name="name" /></form>

http://tomcat.apache.org/

Seriously, I've worked on a whole bunch of PHP sites, and every single one of them had numerous vulnerabilities, primarily SQL injection vulnerabilities. How do you avoid SQL injection attacks? You must use a DB layer that prevents them. PEAR provides protection IF it is used consistently and correctly (sorta like something else that provides protection, maybe the Slashdot crowd doesn't know about it though).

But really, PHP is hostile to the idea of secure code. Every variable can be a string, variables don't need to be declared, a function has no idea what type of input its getting, and most sites don't use proper database layers.

In the Tomcat world, every variable has a type, a variable may not be used before it is declared AND initialized, both these are enforced at COMPILE TIME (something which doesn't exist in PHP), and most sites use reasonable database layers, either PreparedStatements or even better, Hibernate.

I think PHP would be better if sites ran in strict mode (or whatever they call it) but I think few of them do, and to get a legacy codebase to run in strict mode is a major task.

All software should go through a syntax check before it is deployed, a syntax check which checks over every line of code, without waiting for the code to be exercised. That syntax check is often performed by a COMPLIER. Code should be compiled.

Disclaimer: didn't RTFA.
So:
I'm going to be able to access not just applications in my browser, but a whole OS?
Gosh, think of the security implications - where are my files being stored, blahblah.
FUCKING DYNDNS, APACHE & a JAVA VNC CLIENT
Although the rest of the posts seem to suggest that the article actually has sweet F.A. to do with anything.
As an aside, I have heard it mentioned that it is possible to pay for a subscription to slashdot, is the posting of this article some kind of incentive?

Apache Geronimo

For a good logging facility that works for a .Net environment, you can check out
http://logging.apache.org/log4net/

http://logging.apache.org/log4net/ is your friend. it's ported from log4j. it works really well. the documentation is not 100% accurate, so Google is your friend too. we log everything to a single Oracle table, it's nice and easy to mine using SQL statements, it's easy to purge; logging medium is mostly a matter of personal preference. a good logging framework will let you add new loggers / change log levels without touching your app (without even restarting it). BTW, Microsoft's Enterprise Library Logging Block appears to be mostly crap...(http://weblogs.asp.net/lorenh/archive/2005 /02/18/376191.aspx).

I'm unfortunately not too familiar with C#, so I can't comment on it's logging facilities (or lack thereof) other than the .NET EventLog class.

There is a project on Sourceforge called C# Logger that is supposedly similar to log4j in Java. But it seems to be stuck in alpha release mode, and not particularly active.

Just my two cents. Hopefully it helps. :)

A good logging library for .NET is log4net. It is basically the same API as log4j which is a popular logging library for Java.

I also forgot to add...for logging & debugging purposes, I have found log4net http://logging.apache.org/log4net/ to be extremely useful. It is not the easiest logging library to get set up, but it does offer quite a bit of flexibility and is easy to use once you do.

I'd recommend taking a look at Lucene as well.

So, aside from reading books on Information Retrieval and Data Mining, the other easily available reference are open source search engines. In particular, look at the Nutch project, which is actually a pretty high quality search engine implementation. Even better: start contributing to the project.

The OSGi framework mentioned is very cool indeed. It's best known usage is the Eclipse IDE. It can also be used in web applications, where especially the Wicket component web framework delivers a very good integration. There are several users working with OSGi compliant frameworks (most notably Oscar, which is in the Apache incubator under the name Felix), and Wicket. I have used Oscar and Wicket in a commercial product and we were very satisfied with the runtime re-deployment of new components.

AFAIK you don't have to restart tomcat either.
There was some (server.xml?) setting you could enable for hot deploy. But to be fair, I didn't trust that setting to always work back then.

But what Tomcat 5 has (and what I'm sure of that it works) is its manager webapp for (remotely) deploying webapps.
Now, a web interface isn't a nice thing to use for deploying during development but there are ant tasks to automate the whole thing:
Ant Tasks for remote deploying on Tomcat

Looking at the APIs used by the book, I think he forgot two very important ones:

PDFBox - A library that gives you complete parsing control over PDFs. You can create documents from scratch, extract text, merge pages, change text/images/fields, populate forms, etc. With the use of PDFs in corporate environments on the rise, you just can't go wrong with PDFBox in your toolbox.

POI - Does your company use Office documents in any way, shape, or form? Then you need POI. Create spreadsheets and word documents from scratch, modify existing documents, create spreadsheet computers and reports on the fly. The possibiliies are endless, and your boss with thank you for not having to deaL with ANOTHER CSV file.

Most Open Source CMSs aren't much more than blogs or forums on steroids. Very few deal with real content management problems. I still haven't found one that I'm crazy about. I'm trying to work with Apache Lenya right now but it takes a lot of work. Zope/Plone is similar. The power is there for both of them but the initial learning curve is steep.

Oh, and my biggest pet peeve for any CMS site (or any site) is unreadable URLs. It's OK for some applications but for a site where people will be returning to the same page frequently, it should have a sane URL.

There are quite a few Java based open-source CMS like Magnolia (http://www.magnolia.info/en/magnolia.html), Apache Lenya (http://lenya.apache.org/ etc. An exhaustive list of Java based open-source CMS can be find here:
http://java-source.net/open-source/content-managme nt-systems

SpamAssassin & Thunderbird heuristic learning, have been keeping my inbox 99,7% spamfree for the last 2 years.

Stupid as i am, i never realized that i have Microsoft to thank for it.

Did you read the article, or the WHATWG spec?

It specifically mentions:

1. Links with the "ping" attribute should be diffrentiated from other links.
2. There should client-side options to control "ping" behavior, similar to current cookie options: "respond to all", "ignore 3rd party", "ignore all".

FWIW, this really seems dead in the water. First, not too many users will have it enabled (or even available, for that matter). Second, this information is already being reliably collected with cookies, mod_usertrack, javascript, and page redirect tricks -- mostly with no knowledge of the enduser.

Why go with a little-available, easily disable mechanisim when the tried-and-true method is already available?

Admittedly as applied to the original poster's situation that may be so, however as a general point in using the GPL with your development, eg. linking to libraries etc., it is still an obligation that must be cleared. You may not even know what the eventual plans for the code you're writing are - perhaps they're going to resell the system at a later date? In which case you've just created a situation which requires approval first.

"On the other hand, if you use any software that has a EULA, an actual use license, then you are perhaps agreeing to something when you start using it. But I've never seen any open source software with a EULA."

The specific producst mentioned were Sun's JDK (which we'll skip as it's not open-source), Eclipse and Apache. So looking at the final two.

From the Eclipse Public License page:
"THE ACCOMPANYING PROGRAM IS PROVIDED UNDER THE TERMS OF THIS ECLIPSE PUBLIC LICENSE ("AGREEMENT"). ANY USE, REPRODUCTION OR DISTRIBUTION OF THE PROGRAM CONSTITUTES RECIPIENT'S ACCEPTANCE OF THIS AGREEMENT. ".

From the Apache License page:
" TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION"

(emphasis added by me).

So yes, the licenses discussed have terms for use. And the JDK certainly does.

Cheers,
Ian

http://jakarta.apache.org/tapestry/

If you go with Java, there are plenty of other choices than JSF. Struts, while a bit verbose and showing its age, is a very mature framework that scales well and has been used successfully in lots of projects. A lot of people recommend Tapestry or Cocoon. It all depends on the size of your project and what people are experienced with.

A good thing with Java, no matter which framework you choose, is that you have a huge number of open source tools and libs to help you (Eclipse, Netbeans, JUnit, Ant, Maven, CruiseControl, JMeter, PMD, Checkstyle, xdoclet, Hibernate, Spring, Tomcat, commons logging, jsch...) , and there are also plenty of books, online tutorials, and programmers around who know Java.

If you go with Java, there are plenty of other choices than JSF. Struts, while a bit verbose and showing its age, is a very mature framework that scales well and has been used successfully in lots of projects. A lot of people recommend Tapestry or Cocoon. It all depends on the size of your project and what people are experienced with.

A good thing with Java, no matter which framework you choose, is that you have a huge number of open source tools and libs to help you (Eclipse, Netbeans, JUnit, Ant, Maven, CruiseControl, JMeter, PMD, Checkstyle, xdoclet, Hibernate, Spring, Tomcat, commons logging, jsch...) , and there are also plenty of books, online tutorials, and programmers around who know Java.

Both are a good choice if you want to properly engineer a new web-based tool. ASP.NET is probably quicker but if you want to do anything really serious you'll probably want to look at purchasing Visual Studio 2005 rather than just using the Visual Web Developer Express. Also the tool support for JSF isn't nearly as mature so it will probably take longer to implement in JSF than in ASP.NET.

Having said that JSF is still a good choice - particularly if licensing costs and portability are an issue. Apache MyFaces is an excellent framework whose only downside is the poor documentation. JSF can be slower to get started with but I found that it enforces best practices more strictly and once you get the hang of all the XML wiring it wasn't that bad. Another benefit of JSF is that you'll have trouble breaking the MVC pattern but you can pretty easily embed alot of code in ASP.NET unless you properly use code-behind and deliberately seperate out the DAL which isn't the default for the point and click wizards (the DAL separation).

In the end it comes down to a few things. If you have existing C#/VB skills and don't mind being stuck with IIS then go for .NET. If portability is a big issue and you'd really like to run this application on a small server running Jetty(for instance) then go for JSF.

I've had an e-mail address for over 15 years. My spam in the past 2 months is less than I had 10 years ago. I post my main address unobfuscated on /. and 25 other public forums. My signal to noise ratio is 100:1. In 5 days I received about 200 real e-mails and 3 spam.

Me too, me too, me too. Except that 3 spam in 5 days would be unusual for me.

I gave up hosting my own e-mail late last year. I moved all my employees and family to gmail.

Oops, not me too. I still run my own mail on my own domains and have given up nothing. I didn't have to turn over the storage and routing to some distant corporation (even though it's maybe not so distant for me, I work at Google). It's called server side Spamassassin. Mix with sendmail, postfix, or Exim, season with procmail, it's really not that hard. At least not for somebody that was already running mail servers on their domains. There have been good times and bad times in the last 10 years' arms race with spammers, but for now, spam is a solvable problem.

Here are a few things that might help:

http://gmail.google.com/ (webmail with a reasonably good spam filter)
http://spamassassin.apache.org/ (good spam filter for non-web email)
http://bogofilter.sf.net/ (another good spam filter for non-web email)
http://spamcop.net/ (free anti-spam service)
http://spambob.net/ (free receive-only/forwarding/black hole email addresses)

It's not the 20 services the GP promised, of course, but it might help, although experience shows that those who complain the most about spam are also the ones who aren't willing to actually try anything to make it stop, so I'm not sure your cousin will find this useful (your description of a "12-year old airhead" certainly doesn't inspire confidence).

• Wikipedia
  
• Firefox
  
• OpenOffice
  
• BitTorrent
  
• MediaWiki
  
• Xvid
  
• phpBB
  
• Outfoxed
  
• Dyne:bolic
  
• GIMP
  
• Apache
  
• SourceForge
  

As the original author of Tapestry (but not the article on DeveloperWorks, which caught me by surprise) I can say that IBM doesn't have any secret agenda on this. In fact, given that IBM is selling a commercial product that competes head-to-head with Tapestry (their JavaServer Faces, built on top of their WebSphere proprietary Eclipse IDE) it is enlightened of them to cover Tapestry.

Of course, what's going on there is two fold. First, IBM is big enough that different areas of the organization will have different and occasionally competing goals. Primarily, all on-line magazines are constantly hunting for new material to keep the eyeballs looking (and the click rates clicking). IBM doesn't solicit authors to write on particular subjects, they accept existing authors efforts, with the authors pursuing their own interests. Here, Brett happened to be into Tapestry and did a great job providing additional documentation in the form of this article.

I make my living off of Tapestry, so I'm happy to see this kind of coverage, but the framework itself is open source and free, with a very, very liberal license (ASL 2.0). I make money by providing Tapestry support and training. There's your ad.

In even newer news, Tapestry 4.0 final release is now available.

Tapestry is open source and free. It is licensed under the Apache Software License, which allows it to be used even inside proprietary software

I certainly have enough old code laying around to parse enough different config files. There are three advantages to XML. The first is that it is text-based, so you can write it from any programming language that can write text to files, or your favorite text editor. That doesn't set it apart from any other text file format. The second is that there are lots of libraries and tools for dealing with it in any programing language. You don't have to write, debug and maintain tools to parse it. The third is only an advantage if you use it right. That is, the tags and attributes document what your data is. If you do it wrong, your data can end up looking like sloppy ASN.1 littered with angle brackets.

http://httpd.apache.org/

I mostly use KMail at home for the same reasons. Though i use fetchmail to retrieve the mails and procmail to pipe them thru ClamAV and SpamAssassin and finally sort them with some scripts of my own.

The fact Kmail use mail dir format, as mutt, let me also check my mail from a remote ssh session.

Some people might want to have a look to AMaVIS or check SWiK about
- emails
- fetchmail
- procmail
- ClamAV
- SpamAssassin
- KMail (nothing really here)
- mutt