I just read through the language reference, here's my impressions of C#.
Being able to assign the get/set accessors for properties is a real plus. I always liked that about Delphi, and wished Java had a better way of doing that. The indexers are also neat. Nothing really revolutionary, but very useful.
Having primitives "boxed" and "unboxed" in objects as needed is also neat. Save loads and loads of that Hashtable.put("number 1", Integer.parseInt(1)) type code that is just a pain in the ass, without sacrificing the performance of making everything an object all the time. I like that.
I also like the foreach statement, as it does make the language a little more expressive (though nowhere near as expressive as Perl.) I like coding in Java all day, but the code is just so bland.
However, there does seem to be a lot of stuff left over from C/C++ days that is sort of questionable. Like the inclusion of structs and enums. The potential performance benefits of structs are intriguing, but losing much of the OOP nature of your code is not attractive. Enums seem to clutter the language needlessly.
Delegates are interesting, but it seems as if most of the functionality provided by them could also have been accomplished with use of interfaces. Curious however.
I really dislike the whole notion of the unsafe code blocks, and getting around the garbage colelctor when you want to. Why not use hook in a native interface like JNI to allow that sort of code? Even more, since all oject are COM objects, and since you can pull in outside COM objects so easily, that would seem to be a good fit.
The proliferation of modifiers and qualifiers for methos and classes and variables is quite confusing, but then I expected that, since MS has a heritage of that sort of thing in the Win32 API. Particularly odd is the "internal" modifier. I read it is meant to say "used by this project only." Huh? What happened to inheritance and the like?
In the same vein, I don't fully understand the Attribute support, after several readings. It seems as if you could provide your own modifiers and such, sort of a hybrid interface, but they don't seem to document it much either. How odd. Has anyone made sense of this yet?
The preprocessor can be useful sometimes, I agree, but in many cases it just clutters up the code. I'm somewhat disappointed to see this included. I felt Java's lack of preprocessor was a good step forward.
All in all, it's an interesting language that ties together existing principles and cleans up some of the aspects in the process. But as expected, it's nothing new, and since it is trying desperately to hold on to the legions of Win32 C++ programmers, many concessions were made to old-school thinking at the expense of robustness and security.
I agree. Have you looked at the spec for Ant (the new Java-based build tool from Jakarta project of Apache) or the Servlet 2.2 "web app" spec from Sun, or the Tomcat configuration spec from Jakarta/Apache/Sun? Those all use XML for the configuration files, and its a complicated mess.
Its sort of powerful, yes, and it does keep you from having to do the dirty work of actually parsing the file yourself, but it sure makes server administration cumbersome and error prone. Configuring Apache is just so simple: load the httpd.conf, type in a directive or 2, and you're done. With Tomcat, I have to load the XML file and type all the silly tags, which basically means I have to type almost twice as much text to get the same configuration stuff. And then theres the fact that all the tags make the actual configuration parameters hard to read...it's a mess. Ant is the same way. Having it all XML based is sort of a downer, because otherwise, it's a very cool concept and implementation.
XML is probably most useful in communicating between heterogenous services, since it does provide a way to ensure that the data you get makes sense to you. Things like MathML and the like are going to be huge wins. But using XML for ALL text-centric data is just misguided I think. It has significant drawbacks (not the least of which is the speed and size issues.) I think there's going to be a shakeout soon, and something better will come along as a result of some programmer scratching this itch...
I have heard of several corporations that banned the use of.DOC on their network; they make all employees transmit files via network storage/email/whatever in.RTF to eliminate virii and interoperability problems.
Let's face it. Most people just don't need all that shit in their document. Bulleted lists and tables satisfy most people's needs. Whatever happened to optimize for the common case?
Do they make you wear the stiff white shirt of a specific color of computer lore, or are they like lots and lots of tech companies, that recognize being productive means being comfortable (ie, wearing what you like and working how you like?)
Not that this is not intended as a troll...I'm generally interested. I'll be graduating with a BS in CS in 2 years, and I'm trying to get a head start on job prospecting. As such, I don't want to work for a place that imposes arbitrary rules on work habits and the like.
Look at what they're doing with Sun and Java. After working with Sun for 4 years, they refuse to ratify Sun's version of EJB, in effect leaving Sun out to dry. IBM was given access to all of Sun's java IP and this is how they repay them?
I thought this was because Sun was charging licensing fees for thr EJB spec, and IBM didn't think that was fair. Thus, they register a protest by refusing to ratify it.
I'm of the belief that all this build it for the browser hype is going to fall apart sometime soon. Has anyone actually tried to build a real app in DHTML, driven by server-side logic? It's a friggin mess.
Just think about how such apps are built. Client connects to server with a request. Server crunches numbers. Server returns response to client in the form of a page that can be interacted with. Repeat.
Sounds good right? Not really. Look at the system in more detail.
Client/Server interaction is limited to a single request/response. As we all know, every time I sen d something to the server, it has to be explicityly told who I am. It's like your 85 year old grandmother, who can't remember anything about you when you come over to see her.
Second, once client client has their page, they quickly discover its really just a static document, made interactive with a series of ugly hacks. Forms were great to order products, and JavaScript was a good way to ensure your customer entered their address correctly, but combine all that with the layers and CSS-P mess, and suddenly it's like we forgot how to design things well. Suddenly your app can know the position of something, but only if you put it there? Suddenly you have to rewrite a bunch of text to do something simple like move a window across the screen?
The problem with the DHTML platform is that nothing is simple. It doesn't work together in a cohesive fashion. There's a reason why a good, standards compliant, relatively bug-free browser hasn't been built in the last 3 years. The "standard" is too complicated.
Think about why you never really saw the fancy DHTML pages all over the web making everything interactive. The stuff was too complicated to write effectively and robustly, and the browsers couldn't render it anyways.
I really think we are going to see something else designed a little better come along. This platform is just a mess, and more and more people are coming to realize that - more and more people are looking for alternatives.
I think the problem with linking from a legal standpoint is that it is a way to do something very much like hosting files on your site, without actually having the file there. I can put up a 10k page containing links to hundreds of cracks and mp3s and the like, all linked to external sites.
Why is this a problem? For the end user, it's irrelevant that the content (the crack or the mp3) is hosted elsewhere. He went to that one 10k page to get it. It's just as if it WERE on your site.
However, what's the problem with this from a legal standpoint? Well, what does one do about linking to a page that has illegal content, without the author's knowledge? Should an author be required to check the "legality" of all information containted in sites he links to? What about links to links to content?
I think the problem is that there are really two types of links. The first type links to another site, ie, people that post a href="http://www.slashdot.org". The second type posts a link DIRECTLY to a piece of content, ie href="http://download.sourceforge.net/slashcode/sl ash-1.0.4.tar.gz"
While the problems are similar with both types of links (if you can accept that there are two distinct types of links), I have a difficult time justifying linking DIRECTLY to illegal content. ie, if DeCSS is declared illegal, I don't see how you can justify a link to http://www.dvdccaisbad.org/files/decss.zip. That is almost the same as hosting the file yourself, indeed, for the user, it IS the same.
However, links to a site containing illegal content are something that MUST remain legal, for the web to maintain its open and dynamic nature, even if that means illegal content will still be available. It's just not fair to the author of a site to require him to check to ensure http://www.gskp.com/free/ changed it's name from Great Shots of Kids Playing to Gimme Some Kiddie Porn.
However, do we think the law could recognize a distinction between the types of links? I don't think so. The line is too finely drawn, and any rule would either be too broad, and thus have a dampening effect on the web, or would be too narrow, and easily circumvented.
Several times in this thread,/.ers state emphatically that it's "a well-known fact" that MS uses Undocumented APIs to give their software an edge over the competitors software. I am beginning to think that should be "a well-known anecdote." Has anyone ever uncovered conclusive proof of this, that I could be directed to? I find it a difficult claim to swallow without concrete supporting evidence. Is there any to offer?
Given the choice between Q3A and UT, I'll take UT. It's even more polished than Q3A, IMHO.
I find it hard to play UT in a serious setting. The weapons just have have the same feel as Q3A's; by that I mean, they seem much more like a collection of toys that you use to just blast away the player. But Q3A just seems so much more serious and down-to-the-core of deathmatching, you know?
UT's Team Play is arguably better, but only because it includes more gameplay styles out-of-the-box. Q3:Team Arena should fix that. But I still enjoy CTF in Q3A more; the sniper rifle in UT is just painful. Too, too many games turn into boring snipe fests. I find Q3A tends to be a lot more fast-paced, without sacrificing the teamplay aspects.
But still, even after all these years, my all-time favorite deathmatch is Doom II. The addition of the double barrel shotgun was enough to add some speed and flavor to the game without breaking the balance (I think it actually improved upon the balance, as short range kills were now feasible against a skilled BFG user.) The speed was just insane; play Doom II DeathMatch for hours on end and you'll feel like you are high on some illegal substance. I haven't had that same adrenaline rush since...newer games are just too slow.
I hope id brings back the frantic pace in the new Doom.
About a year ago, it looked more and more like id is just going to become a technology company.
Then Q3A came out, and that changed. Q3A was an extremely polished game; one of the most polished I have ever played. The whole thing had a wonderful feel and looked and played consistently. There was a single goal, one driving vision behind the game: provide the best serious, hard-core deathmatch possible. I think it delivered spectacularly.
I look forward to the next Doom. The atmosphere in the original (Doom I episode I in particular) was just EERIE. Updated graphics and a new engine will be amazing.
Remember, Carmack owns a substantial portion of id as well. So they could force him out, just as the owners forced Romero out a few years back (he was a part owner as well.)
Now as to the wisdom of such an action....well, I'll let you draw your own conclusions.
Wolfenstein 2001 IS being worked on; it's called Return to Castle Wolfenstein, being developed by Gray Matter Interactive Studios, which is made up of the former employees of Xatrix (makers of Kingpin.)
The subject line says it all; Metallica has already alienated lots of their fans. Listening to the Napster users' appeals isn't going to change their public image; they have commited a MAJOR screw-up; it's even likely that they realize it, but can't turn back now.
The damage has been done; I can't say I feel sorry for them.
This idea just won't work. It's just too much like the current PC market.
Game consoles are meant to be SIMPLE. Consumer buys console. Consumer buys game, puts CD in, and plays. Done.
PCs, on the other hand, are far from simple. You have to deal issues like: Is my CPU fast enough? Do I have enough RAM? Do I have enough hard drive space? Is my 3D Card fast enough? Is my sound card supported? Is my version of my operating system supported?
That's much too complicatd for consumers to deal with. Having an "open standard" and having multiple implementations of it isn't going to fly in this case; there's simply too much to gain in adding in your own proprietary extensions to differentiate your system from everyone else's. Afterall, why SHOULD the consumer choose me over my competitor, if we are both the same? It just doesn't make sense.
The Lone Gunmen are my least favorite characters on the show. They just sort of babble on in the X-Files equivalent of marketspeak -- "You mean they downloaded the virus into the defense mastrix and how the system is overloaded and can't be rebooted? What about the carrier tones"....
You get the idea. They try to be geeks for non-techy people, but I just don't think it works.
Besides, everytime they come on the plot ends up being "we need to save our friends!" like in that lameass FPS episode.
I really, really hope Fox kills that stupid idea for a Lone Gunmen series after the pilot is made. It just won't be worth watching.
Although, I have to say, both Hook and Abrash have gone on record as saying one of the reasons they left was get to a greener environment. I haven't heard if this was part of Cash's motivation.:)
That I will grant you. Dallas is basically a big field of concrete:)
What worried me about Postgres was that they claimed that before 6.5, there were some sections of code that NOBODY on the development team understood.
I understand that in most large projects, there isn't the one guy that understand all the source. But for any given module of chunk of code, there should be at least one developer (and hopefully many more than that) that understands the code and knows how to maintain it.
Perhaps bumping the developer count to 120 will remedy this; let's sure hope so!
When the "MySQL sucks" article appeared on/. last week, I began to look into Postgres because I began to think that perhaps I really did need some of the ACID features in my database. After 20 minutes of perusing the site, I gave up on the idea. Why?
One of the "highlights" in the developer's page talked about how the new release version 6.5 represents Postgres' mastery over the source. What does that mean? According to the developers, it is the first release where they have someone on the team that understands every section of code.
Let me tell you, that scared me away. I really don't like the idea of using a database where not even the developers have a firm understanding of the source; I don't want a nebulous developer response to a bug or a question. Combine that with the plethora of anecdotes about Postgres being great as long as you don't need speed or reliability (one guy reported having to restart the Postgres daemon every 5 minutes due to memory leaks!)
Perhaps this new arrangement can bring some stability and speed to Postgres. I at least hope they will fix the bugs, as the idea of a free (as in speech) RDBMS is tempting.
Slashdot Mirror
Too bad the friendly HTTP error messages go against the RFC. I thought the RFC said you have to display the error page given to you from the server?
I just read through the language reference, here's my impressions of C#.
Being able to assign the get/set accessors for properties is a real plus. I always liked that about Delphi, and wished Java had a better way of doing that. The indexers are also neat. Nothing really revolutionary, but very useful.
Having primitives "boxed" and "unboxed" in objects as needed is also neat. Save loads and loads of that Hashtable.put("number 1", Integer.parseInt(1)) type code that is just a pain in the ass, without sacrificing the performance of making everything an object all the time. I like that.
I also like the foreach statement, as it does make the language a little more expressive (though nowhere near as expressive as Perl.) I like coding in Java all day, but the code is just so bland.
However, there does seem to be a lot of stuff left over from C/C++ days that is sort of questionable. Like the inclusion of structs and enums. The potential performance benefits of structs are intriguing, but losing much of the OOP nature of your code is not attractive. Enums seem to clutter the language needlessly.
Delegates are interesting, but it seems as if most of the functionality provided by them could also have been accomplished with use of interfaces. Curious however.
I really dislike the whole notion of the unsafe code blocks, and getting around the garbage colelctor when you want to. Why not use hook in a native interface like JNI to allow that sort of code? Even more, since all oject are COM objects, and since you can pull in outside COM objects so easily, that would seem to be a good fit.
The proliferation of modifiers and qualifiers for methos and classes and variables is quite confusing, but then I expected that, since MS has a heritage of that sort of thing in the Win32 API. Particularly odd is the "internal" modifier. I read it is meant to say "used by this project only." Huh? What happened to inheritance and the like?
In the same vein, I don't fully understand the Attribute support, after several readings. It seems as if you could provide your own modifiers and such, sort of a hybrid interface, but they don't seem to document it much either. How odd. Has anyone made sense of this yet?
The preprocessor can be useful sometimes, I agree, but in many cases it just clutters up the code. I'm somewhat disappointed to see this included. I felt Java's lack of preprocessor was a good step forward.
All in all, it's an interesting language that ties together existing principles and cleans up some of the aspects in the process. But as expected, it's nothing new, and since it is trying desperately to hold on to the legions of Win32 C++ programmers, many concessions were made to old-school thinking at the expense of robustness and security.
Anyone else have comments?
I agree. Have you looked at the spec for Ant (the new Java-based build tool from Jakarta project of Apache) or the Servlet 2.2 "web app" spec from Sun, or the Tomcat configuration spec from Jakarta/Apache/Sun? Those all use XML for the configuration files, and its a complicated mess.
Its sort of powerful, yes, and it does keep you from having to do the dirty work of actually parsing the file yourself, but it sure makes server administration cumbersome and error prone. Configuring Apache is just so simple: load the httpd.conf, type in a directive or 2, and you're done. With Tomcat, I have to load the XML file and type all the silly tags, which basically means I have to type almost twice as much text to get the same configuration stuff. And then theres the fact that all the tags make the actual configuration parameters hard to read...it's a mess. Ant is the same way. Having it all XML based is sort of a downer, because otherwise, it's a very cool concept and implementation.
XML is probably most useful in communicating between heterogenous services, since it does provide a way to ensure that the data you get makes sense to you. Things like MathML and the like are going to be huge wins. But using XML for ALL text-centric data is just misguided I think. It has significant drawbacks (not the least of which is the speed and size issues.) I think there's going to be a shakeout soon, and something better will come along as a result of some programmer scratching this itch...
I have heard of several corporations that banned the use of .DOC on their network; they make all employees transmit files via network storage/email/whatever in .RTF to eliminate virii and interoperability problems.
Let's face it. Most people just don't need all that shit in their document. Bulleted lists and tables satisfy most people's needs. Whatever happened to optimize for the common case?
Do they make you wear the stiff white shirt of a specific color of computer lore, or are they like lots and lots of tech companies, that recognize being productive means being comfortable (ie, wearing what you like and working how you like?)
Not that this is not intended as a troll...I'm generally interested. I'll be graduating with a BS in CS in 2 years, and I'm trying to get a head start on job prospecting. As such, I don't want to work for a place that imposes arbitrary rules on work habits and the like.
Look at what they're doing with Sun and Java. After working with Sun for 4 years, they refuse to ratify Sun's version of EJB, in effect leaving Sun out to dry. IBM was given access to all of Sun's java IP and this is how they repay them?
I thought this was because Sun was charging licensing fees for thr EJB spec, and IBM didn't think that was fair. Thus, they register a protest by refusing to ratify it.
Look into Ant, the XML based build tool from Apache.
It's still in its infancy, but the idea is solid I think. Just needs to be developed more.
I'm of the belief that all this build it for the browser hype is going to fall apart sometime soon. Has anyone actually tried to build a real app in DHTML, driven by server-side logic? It's a friggin mess.
Just think about how such apps are built. Client connects to server with a request. Server crunches numbers. Server returns response to client in the form of a page that can be interacted with. Repeat.
Sounds good right? Not really. Look at the system in more detail.
Client/Server interaction is limited to a single request/response. As we all know, every time I sen d something to the server, it has to be explicityly told who I am. It's like your 85 year old grandmother, who can't remember anything about you when you come over to see her.
Second, once client client has their page, they quickly discover its really just a static document, made interactive with a series of ugly hacks. Forms were great to order products, and JavaScript was a good way to ensure your customer entered their address correctly, but combine all that with the layers and CSS-P mess, and suddenly it's like we forgot how to design things well. Suddenly your app can know the position of something, but only if you put it there? Suddenly you have to rewrite a bunch of text to do something simple like move a window across the screen?
The problem with the DHTML platform is that nothing is simple. It doesn't work together in a cohesive fashion. There's a reason why a good, standards compliant, relatively bug-free browser hasn't been built in the last 3 years. The "standard" is too complicated.
Think about why you never really saw the fancy DHTML pages all over the web making everything interactive. The stuff was too complicated to write effectively and robustly, and the browsers couldn't render it anyways.
I really think we are going to see something else designed a little better come along. This platform is just a mess, and more and more people are coming to realize that - more and more people are looking for alternatives.
I think the problem with linking from a legal standpoint is that it is a way to do something very much like hosting files on your site, without actually having the file there. I can put up a 10k page containing links to hundreds of cracks and mp3s and the like, all linked to external sites.
l ash-1.0.4.tar.gz"
Why is this a problem? For the end user, it's irrelevant that the content (the crack or the mp3) is hosted elsewhere. He went to that one 10k page to get it. It's just as if it WERE on your site.
However, what's the problem with this from a legal standpoint? Well, what does one do about linking to a page that has illegal content, without the author's knowledge? Should an author be required to check the "legality" of all information containted in sites he links to? What about links to links to content?
I think the problem is that there are really two types of links. The first type links to another site, ie, people that post a href="http://www.slashdot.org". The second type posts a link DIRECTLY to a piece of content, ie href="http://download.sourceforge.net/slashcode/s
While the problems are similar with both types of links (if you can accept that there are two distinct types of links), I have a difficult time justifying linking DIRECTLY to illegal content. ie, if DeCSS is declared illegal, I don't see how you can justify a link to http://www.dvdccaisbad.org/files/decss.zip. That is almost the same as hosting the file yourself, indeed, for the user, it IS the same.
However, links to a site containing illegal content are something that MUST remain legal, for the web to maintain its open and dynamic nature, even if that means illegal content will still be available. It's just not fair to the author of a site to require him to check to ensure http://www.gskp.com/free/ changed it's name from Great Shots of Kids Playing to Gimme Some Kiddie Porn.
However, do we think the law could recognize a distinction between the types of links? I don't think so. The line is too finely drawn, and any rule would either be too broad, and thus have a dampening effect on the web, or would be too narrow, and easily circumvented.
As such, I think linking must remain legal.
Several times in this thread, /.ers state emphatically that it's "a well-known fact" that MS uses Undocumented APIs to give their software an edge over the competitors software. I am beginning to think that should be "a well-known anecdote." Has anyone ever uncovered conclusive proof of this, that I could be directed to? I find it a difficult claim to swallow without concrete supporting evidence. Is there any to offer?
Given the choice between Q3A and UT, I'll take UT. It's even more polished than Q3A, IMHO.
I find it hard to play UT in a serious setting. The weapons just have have the same feel as Q3A's; by that I mean, they seem much more like a collection of toys that you use to just blast away the player. But Q3A just seems so much more serious and down-to-the-core of deathmatching, you know?
UT's Team Play is arguably better, but only because it includes more gameplay styles out-of-the-box. Q3:Team Arena should fix that. But I still enjoy CTF in Q3A more; the sniper rifle in UT is just painful. Too, too many games turn into boring snipe fests. I find Q3A tends to be a lot more fast-paced, without sacrificing the teamplay aspects.
But still, even after all these years, my all-time favorite deathmatch is Doom II. The addition of the double barrel shotgun was enough to add some speed and flavor to the game without breaking the balance (I think it actually improved upon the balance, as short range kills were now feasible against a skilled BFG user.) The speed was just insane; play Doom II DeathMatch for hours on end and you'll feel like you are high on some illegal substance. I haven't had that same adrenaline rush since...newer games are just too slow.
I hope id brings back the frantic pace in the new Doom.
About a year ago, it looked more and more like id is just going to become a technology company.
Then Q3A came out, and that changed. Q3A was an extremely polished game; one of the most polished I have ever played. The whole thing had a wonderful feel and looked and played consistently. There was a single goal, one driving vision behind the game: provide the best serious, hard-core deathmatch possible. I think it delivered spectacularly.
I look forward to the next Doom. The atmosphere in the original (Doom I episode I in particular) was just EERIE. Updated graphics and a new engine will be amazing.
Remember, Carmack owns a substantial portion of id as well. So they could force him out, just as the owners forced Romero out a few years back (he was a part owner as well.)
Now as to the wisdom of such an action....well, I'll let you draw your own conclusions.
Wolfenstein 2001 IS being worked on; it's called Return to Castle Wolfenstein, being developed by Gray Matter Interactive Studios, which is made up of the former employees of Xatrix (makers of Kingpin.)
He describes how a developer got canned for political reasons and then in the same post says that id is looking to hire.
Move along folks, no conspiracy here.
OF COURSE he is looking to hire. If you just lost a wonderful 3D Artist, wouldn't you want to get another one right away?
The subject line says it all; Metallica has already alienated lots of their fans. Listening to the Napster users' appeals isn't going to change their public image; they have commited a MAJOR screw-up; it's even likely that they realize it, but can't turn back now.
The damage has been done; I can't say I feel sorry for them.
This idea just won't work. It's just too much like the current PC market.
Game consoles are meant to be SIMPLE. Consumer buys console. Consumer buys game, puts CD in, and plays. Done.
PCs, on the other hand, are far from simple. You have to deal issues like: Is my CPU fast enough? Do I have enough RAM? Do I have enough hard drive space? Is my 3D Card fast enough? Is my sound card supported? Is my version of my operating system supported?
That's much too complicatd for consumers to deal with. Having an "open standard" and having multiple implementations of it isn't going to fly in this case; there's simply too much to gain in adding in your own proprietary extensions to differentiate your system from everyone else's. Afterall, why SHOULD the consumer choose me over my competitor, if we are both the same? It just doesn't make sense.
The Lone Gunmen are my least favorite characters on the show. They just sort of babble on in the X-Files equivalent of marketspeak -- "You mean they downloaded the virus into the defense mastrix and how the system is overloaded and can't be rebooted? What about the carrier tones"....
You get the idea. They try to be geeks for non-techy people, but I just don't think it works.
Besides, everytime they come on the plot ends up being "we need to save our friends!" like in that lameass FPS episode.
I really, really hope Fox kills that stupid idea for a Lone Gunmen series after the pilot is made. It just won't be worth watching.
I agree. WHy not Perl? It's infinitely better for text manipulation anyways.
We don't want VB!
I haven't heard an official term, but I think Grand Prarie must be something like the ruptured spleen of Dallas ;)
not that Dallas claims it as such ;)
:)
Got that right. My friends call it the "armpit of dallas."
Although, I have to say, both Hook and Abrash have gone on record as saying one of the reasons they left was get to a greener environment. I haven't heard if this was part of Cash's motivation. :)
:)
That I will grant you. Dallas is basically a big field of concrete
Dallas, TX isn't exactly Nowhereville, Texas.
(And yes, Mesquite is part of the DFW Metroplex).
I believe I didn't explain what I meant clearly.
What worried me about Postgres was that they claimed that before 6.5, there were some sections of code that NOBODY on the development team understood.
I understand that in most large projects, there isn't the one guy that understand all the source. But for any given module of chunk of code, there should be at least one developer (and hopefully many more than that) that understands the code and knows how to maintain it.
Perhaps bumping the developer count to 120 will remedy this; let's sure hope so!
When the "MySQL sucks" article appeared on /. last week, I began to look into Postgres because I began to think that perhaps I really did need some of the ACID features in my database. After 20 minutes of perusing the site, I gave up on the idea. Why?
One of the "highlights" in the developer's page talked about how the new release version 6.5 represents Postgres' mastery over the source. What does that mean? According to the developers, it is the first release where they have someone on the team that understands every section of code.
Let me tell you, that scared me away. I really don't like the idea of using a database where not even the developers have a firm understanding of the source; I don't want a nebulous developer response to a bug or a question. Combine that with the plethora of anecdotes about Postgres being great as long as you don't need speed or reliability (one guy reported having to restart the Postgres daemon every 5 minutes due to memory leaks!)
Perhaps this new arrangement can bring some stability and speed to Postgres. I at least hope they will fix the bugs, as the idea of a free (as in speech) RDBMS is tempting.