If you want all the latest features provided by Microsoft for the SMB/CIFS/AD implementation, then yes by all means use MS servers and pay lots of money for CAL licenses for each server so that you can host everyone, and be prepared to have keep all your workstations in sync with the version on the server because their backwards compatibility sucks.
You are kidding, right? I can probably hookup a W7 with an NT4 domain with less hassle than with samba.
If you want a version that works, continues to work through Windows upgrades, and provides login compatibility with Unix/Linux systems, then use SAMBA; or if you have Linux/Unix clients that you want to authenticate to a Microsoft server - use SAMBA.
This one is hilarious. Do you know that you _need_ to upgrade samba to work seamlessly with the latest Windows versions, right? Do you know that, at each release, you get new bugs and unexpected behaviour? Do you know that some releases are actually unusable (with show-stopper bugs in stuff like remote profiles), right? And that many many times, connectivity problems are "solved" by not using TCP/IP, but by enabling "NetBIOS" (on top of TCP/IP, but deprecated since W2K).
Since Win95, every major release of the OS has had major differences in the SMB/CIFS/AD protocol that means patching older versions to keep compatible (Win95->Win98->WinME->Win2k->WinXP->Vista, or tweaking settings to make the new versions work with the old version (Win7).
Yeah, it's called evolution. Today a W7 machine tries to negotiate an encrypted link, but if you want, with a couple of clicks on the local policy manager, you can downgrade expectations. It works. (Except printing, but that is mostly a "driver" issue)
You run afoul of your own words. Novell did a great job with NetWare and GroupWare for a long time; that is, until Microsoft started doing things to make the login managers for Windows unstable. This pushed a lot of people over to using Microsoft DC/AD solutions as it did not crash Windows.
While I share your respect for Netware, IPX was a shitty, non-routeable protocol, unsuited for modern tcp/ip based connections. Microsoft took over Novell's business not only by providing crappy IPX support, but also by reading the writing on the wall (TCP/IP was the future, and IBM/OS2 were already on board).
in other words, it's no simple job to keep compatibility with Microsoft products as Microsoft actively tries to subvert compatibility.
That is bullshit. Go try a copy of SUN's CIFS deamon (open sourced more than 5 years ago) and then tell me about it. Or check Samba's bugzilla for showstopper bugs. 10 years later, even the support of basic features is somewhat flaky.
Just look at all the various slight (and undocumented) adjustments to the MS Office Binary formats used until MS Office 2007
Show me a complex document format that spans a 20 year period that isn't complex. Go look at the DXF spec (a well known one) and then tell me about it. In contrast, the RTF specfification is clear and concise, even considering Microsoft extensions.
Or look at Microsoft's Kerberos and LDAP implementations used for AD, which have a lot of minor changes from their respective standards so that you can't just pull a standard Kerberos or LDAP implementation out and use it to do AD.
In retrospect, that happened 12 years ago. And given that their Kerberos implementation is *almost* compliant, and that their LDAP implementation is also *almost* compliant, 12 years is a lot of time.
Disclaimer: I actually use Samba as a domain controller on several clients and in our own infrastructure, and I'm deeply familiarized with the pains involved. I've been using Samba (as a server) since 2.1, mostly because it runs on the Unix operating systems I use/deploy, and when comparing to (at least) W2K Server/W2k3, it is dramatically faster serving files.The fact that OpenLDAP is a big stinkin pile of poo doesn't help either - specially considering that - for our infrastructure - using Windows Server is as good as free.
Re:Big shoutout to Tridge and the whole Samba team
on
Samba 4 Enters Beta
·
· Score: 1
I may be wrong, but as far as I recall, SUN's Samba clone is the best non-MS in the market, and kicks Samba ass so far it may be in Australia now:). And yes, it helps that Samba is a big pile of sh**t.
Shhh don't poke holes on unit testing thinggy, you'll wake the zealots! If your models (that's right, I'm using serious buzzwords) aren't easily testable by simple assertions (as any serious model usually isn't), you should change your code so it is more testable! In fact, you probably should write the tests to familiarize yourself with the problem, and _then_ write the actual code. How about that?
No. Your not getting the point. In the Open Source paradigm solutions get implemented by people who are "scratching an itch", so the PHP support is going to be implemented by PHP programmers. Therefore, the fact that the PHP support is of poor quality is a direct indication that the PHP programmers are of poor quality.
Go see your original message. Fact is that Eclipse *does* use PHP, and I'm not talking about the mentioned PHP support. But, then again, while I'm not familiarized with Eclipse internals, I really doubt that the parsing of sourcefiles is made by an external component - the common approach is to implement a generic parser with a set of pluggable dictionaries, so it doesn't really matter if its parsing Java or PHP.
Of course, you haven't followed your own point along its own logical path, which leads to the point that far more than 50% of PHP programmers could not do so.
I see a difference between someone who can write code and a "programmer". I'd expect a good PHP programmer to be as good as a good Java programmer, and equally rare:)
Finally, you also accidentally make a point when you keep using the term programmer rather than software engineer.
It's not innocent. I wouldn't go that far as associating PHP with incompetent engineering, there are cases where it is a good fit - or a requirement. And software engineers don't do "web design". But I find the expression "software engineer" rather silly, as often is used instead of "application lifecycle manager". Very few software products are actually engineered (as in planned in detail in every aspect from start to finish before conception).
Language Design is done by Engineers, not programmers.
More often than not, is done by mathematicians, physicists and the occasional CS major, not engineers.
PHP is a horribly designed language. In fact, I'm being gratuitous when I use the term design at all in that context.
I don't think you'll ever find a PHP programmer/code monkey that disagrees with that. A bit like FTP is a horribly designed protocol, etc. - it exists, many people use it, it gets work done.
To those who had some difficulty reading my original post - I never said advertising is the _only_ source of revenue, did I? I've just stated that the ad business is both the lion share of it, and the link I provided states exactly that. I also said that I doubt that ads at current levels (~3Bn USD) are be enough to keep the business running. I may be wrong, but if I'm right, even a 100% increase in "other" revenue streams won't cover the difference, so how is that so relevant?
Zuckerberg still owns more than 50% of the voting rights for Facebook
The problem is, 50% of a rapidly devaluating asset is half of almost nothing. The freeride of investor's money has ended, and many of those initial investors have already cashed out. If the company isn't clearly turning a profit on the next year or two, the company value will decrease, current bank investments will dry and then the money problems will come. And we all know how Zuckerberg is such a good lider that works well under pressure, so everything will work just fine, right?
This is business - if he starts to get in the way of the investor's interests, you'll see how fast he's replaced because of a <scandal|family problem|insert lame excuse here>.
First google hit - http://www.splatf.com/2012/02/facebook-revenue/
85% of revenue is made from advertising. And those revenue numbers aren't up to par with the company size, or IPO valuation. Not even close.
I'd say that, after the unsurprisingly disappointing IPO, the infinite money faucet has closed. Give them some months to settle, and then you'll start to see less monkeying around new features and more commercial focusing. People didn't buy Facebook shares to "finance the vision". They bought them to make money, and for that, they need to have a business model (and I really doubt that advertising - at current levels - is enough to keep the lights on).
When they start to put more ads, when some new features start to be paid, it will be the beginning of the end.
I actually use NuShpere's PHPEdit (give it a try if you use Windows), but if you think NetBeans is bad, try Eclipse and then tell me about it:D
The latest versions of VS are more close to Eclipse than to NetBeans. That's one of the motives I use SharpDevelop for C# development, instead of VS 2011.
So, I guess I don't know. Is that relevant? So PHP is fine if it's free and/or opensource? Because I've read here a ton of comments of how Java is "superior" in every way, but the Eclipse Foundation website has parts made in PHP? There is no doubt that Eclipse is heavily Java-oriented, so does that mean that they don't even eat their own dogfood?
Naturally, this is what one would expect since any competent developer is going to eschew PHP in the first place.
Oh, you mean a "Java" programmer? We have different ideas of competence. Yes, PHP as a programming language is a big pile of poo, but at least isn't encapsulated poo. People I'd consider "good programmers" usually barf at the idea of both Java and PHP. Programming isn't about APIs and ABIs and what language abstracts better what you are trying to think - programming is about implementing algorithms in restrained environments using instructions that may or may not be machine-oriented. I'd bet you 50% of the Java programmers you know cannot implement simple algorithms they use everyday in a pseudo-language (heaps, garbage collecting, lists, tuples, etc), or explain easily how the OO inheritance works internally. Go ahead, try it and then tell me if I'm wrong.
When you basically make crap up as you go, the IDEs are always going to struggle.
This is hilarious! So I'm a poor programmer because your IDE can't keep up with my code? So, are you saying that if I try to create a project and add eg. the FreeBSD kernel source, my IDE won't grind into a halt? Just with PHP?
And NetBeans, and every other IDE that works well with big PHP projects is wrong, because Eclipse can't be a big pile of poo?
fail. i've done many many web projects in php (don't blame me!) in eclipse with no problem whatsoever. eclipse is a resource hungry beast, but it is manageable and sports features no other ide has.
You didn't specify if your projects were multi-million line PHP projects, or just some wordpress modules. Scale does matter. And manageable is not the same as usable. And resource hungry beast is, to put it simple, an euphemism. And I couldn't care less for extra features I don't use. The mentality of "one size fits all" is very Java-like, people that can't conceive any kind of application without OO concepts, but usually fail to understand basic programming concepts. Maybe that's not your case, but the comment someone made on your comment is the poster for it.
So, parsing textfiles isn't really parsing textfiles, and Java source has special magic that only Eclipse can see (since every other multi-language IDE seems to have no problem with it)?
And your post wouldn't be completely hilarious if - apparently - the Eclipse website resorts to PHP! With urls similar to http://www.eclipse.org/projects/listofprojects.php, it's not like they don't know PHP exists.
Is multithreading absolutely necessary for every application? Well, let's see - you have full blown SGBDs without mulithreading. Can some of those "multithreaded" applications be re-tailored using an event-driven approach? Probably. Can most of that threaded work be emulated by run queues? Shure.
Or a lack of interactive debugging.
While interactive debugging is nice, it's not a killer feature. And it's not like there really isn't one available (http://www.php-debugger.com/dbg/)
Or the absence of support for modules.
Define "modules". Because I shure can use "modules" - made in PHP code. Binary modules need installation privileges, but it is quite easy not only to add them, but to build them from 3rd party libraries.
Or a crazy type conversion system
Hey, no argument there.
Or an inconsistent and poorly-designed standard library
It's a work-in-progress. It's getting better, and frameworks help. Smart IDEs also help (do you know the order of every parameter in every C standard library?) It is not enough, but it's a big help.
Why not just choose a platform that doesn't have these issues?
You mean, a platform where I need to wait between 15s and 2 minutes until I can actually run a binary? Or a platform that, by having step-by-step debugging promotes laziness (yeah I know, hilarious, right?)? Or a platform that insists on shoving down my throat their way of doing things, and for simple stuff I'd have to instantiate objects like there's no tomorrow.
That some people have got decent results out of PHP does not mean its a good language.
It means it's good enough to get the job done. At the end of the day, that is what matters - shipping good and functional product. And usually maintainable, too.
That doesn’t mean anything. A good carpenter can drive in a nail with either a rock or a hammer, but how many carpenters do you see bashing stuff with rocks? Part of what makes a good developer is the ability to choose the tools that work best.
Stick with programming, and leave both carpentry and metaphores to the professionals. At least, regarding programming, you raise some valid points.
As kind of a disclaimer, I have a strong background in assembly (x86). New instructions in almost every cpu. No "typed variables". No objects (but I used to do OO programming in asm). No native support for threading (new generation Intel has some helpful instructions). 500 different ways of doing stuff. No standard library, and the operating system APIs aren't that consistent. You do have interactive debugging, and your app will at least run in native speed. Oh and when I used asm as my main language, I usually had a high level of productivity, so it wasn't really "a rock".
While I do understand your point (and agree to a certain extent), frameworks usually provide you a nice set of consistent components that have been used and TESTED by a ton of people before you. Shure there are bugs, and for simpler applications they may be overkill, but you also get the benefit of a (mostly) tried-and-true library. If your task is designing an application, it makes no sense wasting your time developing, testing and debugging every single core funcionality you need (eg. database api, routing, caching, locale handling, etc). If even the code produced by some (very smart) framework programmers has bugs after extensive testing and usage by third parties, imagine your own code.
It is nice to see someone stand by their opinion, mr Anonymous Coward.
It is amazing how the world hasn't collapsed yet (given the amount of PHP products and the amount of PHP code deployed) by looking at that list. If PHP programmers can work around all those defects and limitations and create applications that work, I wouldn't call them hacks. They must be wizards, right? Or maybe they just try not to think of it as Java and have a coding style that will avoid most of the defects mentioned. Maybe.
I find it funny that you think PHP is garbage (granted, most of the nitpicks are true, but nobody is forcing you to use it), but somewhat mention Eclipse as a good thing. From my (limited) experience, Eclipse is a poor excuse of an IDE that can't even handle gracefuly a PHP application such as Magento (a somewhat big PHP codebase, but small if compared to an "enterprise" Java project).
The notion of "human readable" varies from human to human. I have a friend that *loves* haskell, but for me the code looks like a corrupted text file. On the other hand, I've identified and corrected errors in programs made with high-level languages from an assembly dump of the code. What I find simple and easy to read may not be that easy to you, or vice-versa.
Oh and I do work with PHP, and I hate the ambiguity of the code.
What you describe sounds a lot like many different existing applications. You may benefit from checking out what is already out there (opensource) and see if it is a good fit (or a good starting point) instead of building your own. Also, you may get some new ideas for your own design, so you won't waste your time.
Regarding programming languages, your decision may be conditioned to the deployment options. Do you want to run it on your own infrastructure, or do you want to put it on a shared host/cloud provider? For shared hosts, PHP or Python may be a good option.
Other aspect to consider is the widget toolkit. Do you have experience with one (or several) that may be more suited to a given language? As an example, GWT is Java-oriented, Dojo integrates well with Zend Framework (PHP), and (AFAIK) both ExtJS and DHTMLX are more language-agnostic. I actually use a lot DHTMLX and their scheduler component may be a good fit for what you're trying to do. Also, (at least) both DHTMLX and ExtJS have a design tool, so you can build your interface without the need for programming or any server-side code.
Finally, do you have other specific requirements, such as scalability, SGBD to use, multi-language support, big persistent data, complex objects, etc? That may also influence the choice of both the language and the framework. Remember, PHP applications are (mostly) stateless, and at every request your application starts from scratch. You can use cache and other tricks, but it will take you only so far - and if you are used to Java, you may take some time to adapt to these limitations.
My personal choice would be PHP with Zend Framework (v1.x), but I don't really like Java and don't have that much experience with other fancy web-oriented languages. Zend Framework is quite complete and probably will give you all the funcionality you may need for your application.
Oracle has no way of "taking back" the already available source. The only problem is that any new ZFS features implemented by the FreeBSD team will probably be incompatible with the Solaris version. But, looking at v28, you may have enough funcionality for a decade (there is encryption missing, and I think dump still not works as expected, but that's about it).
If you want a BSD-licensed alternative, try DragonFly BSD. It comes with the HAMMER filesystem, that provides some (most?) of ZFS funcionality.
Yeah, I know. The situations when the logs aren't that helpful are few enough for me to rembember the details, but the point is that they happen. The quality of the log isn't directly tied to the platform, but to how applications are designed. It's not because you run Linux that messages for a given application will automagically improve.
You must be writing all your software yourself, and auditing all third-party source before you compile it in your audited (or self-written) compiler as it seems you don't see any reason to trust anyone.
I just don't assume that the software is trustable just because some magical Canonical fairies compiled it, and the connection to their server is encrypted. How is this better than windows? (and that was the discussion)
As a sidenote, I usually build everything from ports in my servers (and the base system is rebuilt from a direct copy from the development source). While I cannot audit/control the source, I get a pretty good idea of what is actually installed and the dependencies of the most common applications I use. If a given application is perceived as untrusted, I can always rely on the available methods to try to isolate it as much as possible (securelevels, MAC, chroots, jails, systrace, capiscum, etc). It's not like there is no middle ground between thinking that a repository is farily safe and writing your own software from scratch.
Not as skilled as you seem to think, which simply further reinforces the idea that you're none too bright. As TheLink pointed out here [slashdot.org], you're probably just not enabling any verbosity in your logging. I'm even willing to bet you thought you were all kinds of smart, and disabled the default logging level, figuring it would save you some space on/var/log
Or you're just assuming stuff out of your ass. How about you test it and then tell me if I'm wrong?
Please don't treat me as a retard. I do have the habit of reading the documentation, and at level 3/4 you should see any certificate error (as you see, if eg. the cert isn't valid for the given CA or has expired).
Slashdot Mirror
If you want all the latest features provided by Microsoft for the SMB/CIFS/AD implementation, then yes by all means use MS servers and pay lots of money for CAL licenses for each server so that you can host everyone, and be prepared to have keep all your workstations in sync with the version on the server because their backwards compatibility sucks.
You are kidding, right? I can probably hookup a W7 with an NT4 domain with less hassle than with samba.
If you want a version that works, continues to work through Windows upgrades, and provides login compatibility with Unix/Linux systems, then use SAMBA; or if you have Linux/Unix clients that you want to authenticate to a Microsoft server - use SAMBA.
This one is hilarious. Do you know that you _need_ to upgrade samba to work seamlessly with the latest Windows versions, right? Do you know that, at each release, you get new bugs and unexpected behaviour? Do you know that some releases are actually unusable (with show-stopper bugs in stuff like remote profiles), right? And that many many times, connectivity problems are "solved" by not using TCP/IP, but by enabling "NetBIOS" (on top of TCP/IP, but deprecated since W2K).
Since Win95, every major release of the OS has had major differences in the SMB/CIFS/AD protocol that means patching older versions to keep compatible (Win95->Win98->WinME->Win2k->WinXP->Vista, or tweaking settings to make the new versions work with the old version (Win7).
Yeah, it's called evolution. Today a W7 machine tries to negotiate an encrypted link, but if you want, with a couple of clicks on the local policy manager, you can downgrade expectations. It works. (Except printing, but that is mostly a "driver" issue)
You run afoul of your own words. Novell did a great job with NetWare and GroupWare for a long time; that is, until Microsoft started doing things to make the login managers for Windows unstable. This pushed a lot of people over to using Microsoft DC/AD solutions as it did not crash Windows.
While I share your respect for Netware, IPX was a shitty, non-routeable protocol, unsuited for modern tcp/ip based connections. Microsoft took over Novell's business not only by providing crappy IPX support, but also by reading the writing on the wall (TCP/IP was the future, and IBM/OS2 were already on board).
in other words, it's no simple job to keep compatibility with Microsoft products as Microsoft actively tries to subvert compatibility.
That is bullshit. Go try a copy of SUN's CIFS deamon (open sourced more than 5 years ago) and then tell me about it. Or check Samba's bugzilla for showstopper bugs. 10 years later, even the support of basic features is somewhat flaky.
Just look at all the various slight (and undocumented) adjustments to the MS Office Binary formats used until MS Office 2007
Show me a complex document format that spans a 20 year period that isn't complex. Go look at the DXF spec (a well known one) and then tell me about it. In contrast, the RTF specfification is clear and concise, even considering Microsoft extensions.
Or look at Microsoft's Kerberos and LDAP implementations used for AD, which have a lot of minor changes from their respective standards so that you can't just pull a standard Kerberos or LDAP implementation out and use it to do AD.
In retrospect, that happened 12 years ago. And given that their Kerberos implementation is *almost* compliant, and that their LDAP implementation is also *almost* compliant, 12 years is a lot of time.
Disclaimer: I actually use Samba as a domain controller on several clients and in our own infrastructure, and I'm deeply familiarized with the pains involved. I've been using Samba (as a server) since 2.1, mostly because it runs on the Unix operating systems I use/deploy, and when comparing to (at least) W2K Server/W2k3, it is dramatically faster serving files.The fact that OpenLDAP is a big stinkin pile of poo doesn't help either - specially considering that - for our infrastructure - using Windows Server is as good as free.
I may be wrong, but as far as I recall, SUN's Samba clone is the best non-MS in the market, and kicks Samba ass so far it may be in Australia now :). And yes, it helps that Samba is a big pile of sh**t.
Shhh don't poke holes on unit testing thinggy, you'll wake the zealots! If your models (that's right, I'm using serious buzzwords) aren't easily testable by simple assertions (as any serious model usually isn't), you should change your code so it is more testable! In fact, you probably should write the tests to familiarize yourself with the problem, and _then_ write the actual code. How about that?
No. Your not getting the point. In the Open Source paradigm solutions get implemented by people who are "scratching an itch", so the PHP support is going to be implemented by PHP programmers. Therefore, the fact that the PHP support is of poor quality is a direct indication that the PHP programmers are of poor quality.
Go see your original message. Fact is that Eclipse *does* use PHP, and I'm not talking about the mentioned PHP support. But, then again, while I'm not familiarized with Eclipse internals, I really doubt that the parsing of sourcefiles is made by an external component - the common approach is to implement a generic parser with a set of pluggable dictionaries, so it doesn't really matter if its parsing Java or PHP.
Of course, you haven't followed your own point along its own logical path, which leads to the point that far more than 50% of PHP programmers could not do so.
I see a difference between someone who can write code and a "programmer". I'd expect a good PHP programmer to be as good as a good Java programmer, and equally rare :)
Finally, you also accidentally make a point when you keep using the term programmer rather than software engineer.
It's not innocent. I wouldn't go that far as associating PHP with incompetent engineering, there are cases where it is a good fit - or a requirement. And software engineers don't do "web design". But I find the expression "software engineer" rather silly, as often is used instead of "application lifecycle manager". Very few software products are actually engineered (as in planned in detail in every aspect from start to finish before conception).
Language Design is done by Engineers, not programmers.
More often than not, is done by mathematicians, physicists and the occasional CS major, not engineers.
PHP is a horribly designed language. In fact, I'm being gratuitous when I use the term design at all in that context.
I don't think you'll ever find a PHP programmer/code monkey that disagrees with that. A bit like FTP is a horribly designed protocol, etc. - it exists, many people use it, it gets work done.
To those who had some difficulty reading my original post - I never said advertising is the _only_ source of revenue, did I? I've just stated that the ad business is both the lion share of it, and the link I provided states exactly that. I also said that I doubt that ads at current levels (~3Bn USD) are be enough to keep the business running. I may be wrong, but if I'm right, even a 100% increase in "other" revenue streams won't cover the difference, so how is that so relevant?
Zuckerberg still owns more than 50% of the voting rights for Facebook
The problem is, 50% of a rapidly devaluating asset is half of almost nothing. The freeride of investor's money has ended, and many of those initial investors have already cashed out. If the company isn't clearly turning a profit on the next year or two, the company value will decrease, current bank investments will dry and then the money problems will come. And we all know how Zuckerberg is such a good lider that works well under pressure, so everything will work just fine, right?
This is business - if he starts to get in the way of the investor's interests, you'll see how fast he's replaced because of a <scandal|family problem|insert lame excuse here>.
First google hit - http://www.splatf.com/2012/02/facebook-revenue/
85% of revenue is made from advertising. And those revenue numbers aren't up to par with the company size, or IPO valuation. Not even close.
I'd say that, after the unsurprisingly disappointing IPO, the infinite money faucet has closed. Give them some months to settle, and then you'll start to see less monkeying around new features and more commercial focusing. People didn't buy Facebook shares to "finance the vision". They bought them to make money, and for that, they need to have a business model (and I really doubt that advertising - at current levels - is enough to keep the lights on).
When they start to put more ads, when some new features start to be paid, it will be the beginning of the end.
I actually use NuShpere's PHPEdit (give it a try if you use Windows), but if you think NetBeans is bad, try Eclipse and then tell me about it :D
The latest versions of VS are more close to Eclipse than to NetBeans. That's one of the motives I use SharpDevelop for C# development, instead of VS 2011.
I guess you don't realize how Open Source works.
So, I guess I don't know. Is that relevant? So PHP is fine if it's free and/or opensource? Because I've read here a ton of comments of how Java is "superior" in every way, but the Eclipse Foundation website has parts made in PHP? There is no doubt that Eclipse is heavily Java-oriented, so does that mean that they don't even eat their own dogfood?
Naturally, this is what one would expect since any competent developer is going to eschew PHP in the first place.
Oh, you mean a "Java" programmer? We have different ideas of competence. Yes, PHP as a programming language is a big pile of poo, but at least isn't encapsulated poo. People I'd consider "good programmers" usually barf at the idea of both Java and PHP. Programming isn't about APIs and ABIs and what language abstracts better what you are trying to think - programming is about implementing algorithms in restrained environments using instructions that may or may not be machine-oriented. I'd bet you 50% of the Java programmers you know cannot implement simple algorithms they use everyday in a pseudo-language (heaps, garbage collecting, lists, tuples, etc), or explain easily how the OO inheritance works internally. Go ahead, try it and then tell me if I'm wrong.
When you basically make crap up as you go, the IDEs are always going to struggle.
This is hilarious! So I'm a poor programmer because your IDE can't keep up with my code? So, are you saying that if I try to create a project and add eg. the FreeBSD kernel source, my IDE won't grind into a halt? Just with PHP?
And NetBeans, and every other IDE that works well with big PHP projects is wrong, because Eclipse can't be a big pile of poo?
says it all.
Says it? http://www.eclipse.org/projects/listofprojects.php. Well, aparently even Eclipse resorts to PHP.
fail. i've done many many web projects in php (don't blame me!) in eclipse with no problem whatsoever. eclipse is a resource hungry beast, but it is manageable and sports features no other ide has.
You didn't specify if your projects were multi-million line PHP projects, or just some wordpress modules. Scale does matter. And manageable is not the same as usable. And resource hungry beast is, to put it simple, an euphemism. And I couldn't care less for extra features I don't use. The mentality of "one size fits all" is very Java-like, people that can't conceive any kind of application without OO concepts, but usually fail to understand basic programming concepts. Maybe that's not your case, but the comment someone made on your comment is the poster for it.
Yes, because throwing up buzzwords is what programming is all about.
So, parsing textfiles isn't really parsing textfiles, and Java source has special magic that only Eclipse can see (since every other multi-language IDE seems to have no problem with it)?
And your post wouldn't be completely hilarious if - apparently - the Eclipse website resorts to PHP! With urls similar to http://www.eclipse.org/projects/listofprojects.php, it's not like they don't know PHP exists.
"avoid" support for multithreading
Is multithreading absolutely necessary for every application? Well, let's see - you have full blown SGBDs without mulithreading. Can some of those "multithreaded" applications be re-tailored using an event-driven approach? Probably. Can most of that threaded work be emulated by run queues? Shure.
Or a lack of interactive debugging.
While interactive debugging is nice, it's not a killer feature. And it's not like there really isn't one available (http://www.php-debugger.com/dbg/)
Or the absence of support for modules.
Define "modules". Because I shure can use "modules" - made in PHP code. Binary modules need installation privileges, but it is quite easy not only to add them, but to build them from 3rd party libraries.
Or a crazy type conversion system
Hey, no argument there.
Or an inconsistent and poorly-designed standard library
It's a work-in-progress. It's getting better, and frameworks help. Smart IDEs also help (do you know the order of every parameter in every C standard library?) It is not enough, but it's a big help.
Why not just choose a platform that doesn't have these issues?
You mean, a platform where I need to wait between 15s and 2 minutes until I can actually run a binary? Or a platform that, by having step-by-step debugging promotes laziness (yeah I know, hilarious, right?)? Or a platform that insists on shoving down my throat their way of doing things, and for simple stuff I'd have to instantiate objects like there's no tomorrow.
That some people have got decent results out of PHP does not mean its a good language.
It means it's good enough to get the job done. At the end of the day, that is what matters - shipping good and functional product. And usually maintainable, too.
That doesn’t mean anything. A good carpenter can drive in a nail with either a rock or a hammer, but how many carpenters do you see bashing stuff with rocks? Part of what makes a good developer is the ability to choose the tools that work best.
Stick with programming, and leave both carpentry and metaphores to the professionals. At least, regarding programming, you raise some valid points.
As kind of a disclaimer, I have a strong background in assembly (x86). New instructions in almost every cpu. No "typed variables". No objects (but I used to do OO programming in asm). No native support for threading (new generation Intel has some helpful instructions). 500 different ways of doing stuff. No standard library, and the operating system APIs aren't that consistent. You do have interactive debugging, and your app will at least run in native speed. Oh and when I used asm as my main language, I usually had a high level of productivity, so it wasn't really "a rock".
While I do understand your point (and agree to a certain extent), frameworks usually provide you a nice set of consistent components that have been used and TESTED by a ton of people before you. Shure there are bugs, and for simpler applications they may be overkill, but you also get the benefit of a (mostly) tried-and-true library. If your task is designing an application, it makes no sense wasting your time developing, testing and debugging every single core funcionality you need (eg. database api, routing, caching, locale handling, etc). If even the code produced by some (very smart) framework programmers has bugs after extensive testing and usage by third parties, imagine your own code.
It is nice to see someone stand by their opinion, mr Anonymous Coward.
It is amazing how the world hasn't collapsed yet (given the amount of PHP products and the amount of PHP code deployed) by looking at that list. If PHP programmers can work around all those defects and limitations and create applications that work, I wouldn't call them hacks. They must be wizards, right? Or maybe they just try not to think of it as Java and have a coding style that will avoid most of the defects mentioned. Maybe.
I find it funny that you think PHP is garbage (granted, most of the nitpicks are true, but nobody is forcing you to use it), but somewhat mention Eclipse as a good thing. From my (limited) experience, Eclipse is a poor excuse of an IDE that can't even handle gracefuly a PHP application such as Magento (a somewhat big PHP codebase, but small if compared to an "enterprise" Java project).
The notion of "human readable" varies from human to human. I have a friend that *loves* haskell, but for me the code looks like a corrupted text file. On the other hand, I've identified and corrected errors in programs made with high-level languages from an assembly dump of the code. What I find simple and easy to read may not be that easy to you, or vice-versa.
Oh and I do work with PHP, and I hate the ambiguity of the code.
What you describe sounds a lot like many different existing applications. You may benefit from checking out what is already out there (opensource) and see if it is a good fit (or a good starting point) instead of building your own. Also, you may get some new ideas for your own design, so you won't waste your time.
Regarding programming languages, your decision may be conditioned to the deployment options. Do you want to run it on your own infrastructure, or do you want to put it on a shared host/cloud provider? For shared hosts, PHP or Python may be a good option.
Other aspect to consider is the widget toolkit. Do you have experience with one (or several) that may be more suited to a given language? As an example, GWT is Java-oriented, Dojo integrates well with Zend Framework (PHP), and (AFAIK) both ExtJS and DHTMLX are more language-agnostic. I actually use a lot DHTMLX and their scheduler component may be a good fit for what you're trying to do. Also, (at least) both DHTMLX and ExtJS have a design tool, so you can build your interface without the need for programming or any server-side code.
Finally, do you have other specific requirements, such as scalability, SGBD to use, multi-language support, big persistent data, complex objects, etc? That may also influence the choice of both the language and the framework. Remember, PHP applications are (mostly) stateless, and at every request your application starts from scratch. You can use cache and other tricks, but it will take you only so far - and if you are used to Java, you may take some time to adapt to these limitations.
My personal choice would be PHP with Zend Framework (v1.x), but I don't really like Java and don't have that much experience with other fancy web-oriented languages. Zend Framework is quite complete and probably will give you all the funcionality you may need for your application.
Oracle has no way of "taking back" the already available source. The only problem is that any new ZFS features implemented by the FreeBSD team will probably be incompatible with the Solaris version. But, looking at v28, you may have enough funcionality for a decade (there is encryption missing, and I think dump still not works as expected, but that's about it).
If you want a BSD-licensed alternative, try DragonFly BSD. It comes with the HAMMER filesystem, that provides some (most?) of ZFS funcionality.
Yeah, I know. The situations when the logs aren't that helpful are few enough for me to rembember the details, but the point is that they happen. The quality of the log isn't directly tied to the platform, but to how applications are designed. It's not because you run Linux that messages for a given application will automagically improve.
You must be writing all your software yourself, and auditing all third-party source before you compile it in your audited (or self-written) compiler as it seems you don't see any reason to trust anyone.
I just don't assume that the software is trustable just because some magical Canonical fairies compiled it, and the connection to their server is encrypted. How is this better than windows? (and that was the discussion)
As a sidenote, I usually build everything from ports in my servers (and the base system is rebuilt from a direct copy from the development source). While I cannot audit/control the source, I get a pretty good idea of what is actually installed and the dependencies of the most common applications I use. If a given application is perceived as untrusted, I can always rely on the available methods to try to isolate it as much as possible (securelevels, MAC, chroots, jails, systrace, capiscum, etc). It's not like there is no middle ground between thinking that a repository is farily safe and writing your own software from scratch.
Not as skilled as you seem to think, which simply further reinforces the idea that you're none too bright. As TheLink pointed out here [slashdot.org], you're probably just not enabling any verbosity in your logging. I'm even willing to bet you thought you were all kinds of smart, and disabled the default logging level, figuring it would save you some space on /var/log
Or you're just assuming stuff out of your ass. How about you test it and then tell me if I'm wrong?
Please don't treat me as a retard. I do have the habit of reading the documentation, and at level 3/4 you should see any certificate error (as you see, if eg. the cert isn't valid for the given CA or has expired).