Slashdot Mirror
What's Wrong With the FOSS Community?
An anonymous reader writes "Patrick McFarland, one of the major Free Software Magazine authors, has completed his second article on whats wrong with the Free/Open Source Software (FOSS) community, and what we face in this world. He touches on ESR's Cathedral and the Bazaar essay briefly, and warns against cherry-picking style software development."
I'm pissed. I'm disgusted. Santa is in dire need of some improvements. When he brings all the presents, he does it in the middle of the night, so I have to wait until the morning to get them. When I do get up in the morning to see what I got, all that trashy wrapping paper is in the way, delaying my enjoyment.
Also he drinks all the milk and eats all the cookies!
Good leader > no leader >> bad leader
Nothing in this piece convinces that common sense is wrong.
There is nothing wrong with the FOSS community, however there are a small number of very vocal people who are total assholes towards people new to things such as Linux. I am not a Linux n00b as I have been using it on and off since 1996/7 however when I first gave Gentoo a try (back in 2004 i believe) all I got was abuse when I asked for help with some things. There are a small number of groups within the FOSS community who give it a bad name, however this is the same with most communities IMHO. Ubuntu are doing a lot of good not just with their decent distribution but with a positive and helpful community as well. Infact this is probably the best thing about Ubuntu.
When did the 'FOSS community' become an entity that could be analyzed as a single group so that you could point at it saying that's what's wrong with it?
"I use a Mac because I'm just better than you are."
I kind of dislike the Open Source fanosophy... Sorry, philosophy, but that article was a waste of bandwidth. Of course, in free/Open Source everyone does as they want. Yes, it's a Bazaar, but that's the way it's suposed to be. I do whatever i want in my freetime, but I must do whatever I'm told at work. And that is not going to change.And that has been so since the creation or the evolution from monkeys. And the world has not ended because of that.
It's time to realise that Abble's products are the biggest abomination these days. Just say NO to the dumb iAbble way!!
That one thing is that FOSS can not be the end all and be all of software.
Not every software need will be be solved with FOSS.
There needs to be freedom to write Open and Closed source software. That is what bugs me are people that think selling a closed source package is evil. I just don't think that the FOSS model can work for every program.
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
Clearly, when ideological differences get in the way of even naming the community, you have a problem. Then again, having a common enemy will never be enough to guarantee harmony.
From the article: :-) ) until it matures and shines. The linux kernel is a wonderful example.
Unlike in the Cathedral, the Bazaar has no official leadership.
Isn't this what enables FOSS? Most of the FOSS don't have official leadership (other than the creator of course
I read the article, and I still couldn't tell you what it says. It talks about bazaar, and Gnome and development, but it has no content! I don't think it said *anything*. From the book: Harmless.
I challenge thee to summariser it.
This is what (/usr/bin/ots) a text summariser said (interesting to note it tents to focus on cathedral-style, bazaar-style, and gnome bashing)
A few years back, Eric S. Entitled The Cathedral and the Bazaar, he wrote about how the Free/Open Source Software (FOSS) community does what it wants when it wants to. In Cathedral-style projects, your not-so-friendly neighborhood PHB (fueled by the lies from various ugly hunch-backed minions), although wrong 120% of the time, says what goes in a project. Backed by the Free Software Foundation and the FOSS community as a whole, the GNOME project for many years just added lots and lots of feature creep and otherwise unnamed bloat.
The GNOME project lacked true vision for those years, and feature creep and other long term development problems rushed in to fill that hole. Problem is, many projects are just like GNOME. Incidentally, few Cathedral-style projects suffer from lack of vision: those that do simply die off and are never heard from again. Bazaar-style development allows projects to be in a zombie state for long periods of time, where it is vastly expensive for a Cathedral-style project to do the same. someone with vision (corrupt or not) would control a project, driving development behind it, and have the project reach goals in specific time frames.
Another thing that's wrong with the "community" is writing an article detailing what's wrong with the community and then bashing a project like GNOME, which for all its failings does what it needs to do, is very much active and has a large user and developer following. So I guess this guy must be a "KDE fanboy"... and so it goes.
Web2.0: I love when people Flickr my cuil and digg my boingboing until my google is reddit and I start to yahoo
The biggest problem with the FOSS community is its tolerance for whiny fuckers who can't understand that we do this for fun and you have absolutely no right to complain about something you got for free.
How we know is more important than what we know.
> Not only did Dawes lack vision, he got in the
> way of everyone who did have vision.
That's rather well said. If you're the author of a successful open source project and you find yourself unable to keep working on it, do you have a duty to turn it over to the other developers for continued maintenance? I can't think of a reason not to, and if you don't, it'll either die or get forked, both of which aren't pleasant outcomes.
The Army reading list
Been a Unix/Linux admin for eight years. Been running slackware on laptop as my sole OS for the past five years. I've seen a lot of changes. It's never been better. Sure it's a bazaar, but isn't that how it's supposed to be? Hey, if you don't like gnome, choose something else among the dozens of choices out there.
Perhaps the real problem is the plethora of side-liners, pundits, philosophers, and magazine authors who have nothing better to do than sit around and draw erroneous conclusions. I call these people OSS arm-chair experts. We don't need 'em. Seems the people with most to say write the least amount of code. Maybe they should learn to program and get involved rather than digging too deeply into what's wrong. Be positive.
Someday pigs will fly
Someday hell will freeze over
Someday bears will be catholic and popes will shit in the woods
Someday poeple will stop using weasel words
the foss community is a mirror of the world in general. the people in it are no different to those in any other community. i've met the most generous and helpful people, and also the most nasty. there's nothing "wrong" with the FOSS people, it's just human nature. right and wrong are just a matter of opinion.
If you mod me down, I will become more powerful than you can imagine....
The major difference between FOSS and other communities are that the people in a FOSS community share far fewer specific goals than other communities. Some people want something fixed **now**. Others want it fixed **properly**, no matter how long that takes. Others just piss and moan.
Engineering is the art of compromise.
Though I may not agree with some of his details, overall he's spot on in his argument. He ascertains that the FOSS community - when lacking Cathedral-like leadership - will suffer and potentially flounder. Using the example of x.11 / x.org, he correctly summarizes what is a partial issue with FOSS.
HOWEVER - I think it is very good that such a review exists. As the benevolent dictator of my staff, I encourage ideas and help move software projects forward. I can learn from the FOSS community and their mistakes.
I certainly hope that the kernel development and many other such projects (KDE) follow this type of path.
The Kai's Semi-Updated Website Thingy
But they aren't selling anything ... FOSS developers code because it's fun; that's their compensation, not money. It's no excuse to be an ass, but I don't really see why they should necessarily cater to anyone that isn't contributing in a tangible way.
One big problem that I see if the violence inherent in the community. Everything is a "war" or a "Battle". An "OS War", a "browser war". This article is titled "inside the mind of the enemy". Community != War. If I didn't know better, I'd think that the whole OSS movement was being led by our own current war-loving government (war on drugs, war on terrorism). How about dropping the hostility, for starters?
Well, for those people who don't do this for fun, and instead do this because they want to show the world how great they are or something, you can complain as much as you like. "I'm great." "No you're not, look at this fuck up." "What do you want for nothing?" doesn't follow, I conceed that.
How we know is more important than what we know.
Yep, the boring stuff doesn't get done unless there's incentive to do.
A leader without the ability to fire someone or give them a pay raise isn't going to be able to provide much incentive.
But with FOSS, I (the end user) can email the coder and offer to pay him/her to finish a feature I'd like or do some other boring job. And that is one of the great things about FOSS. Once I pay for it, everyone benefits from it (including me).
Try doing that with closed source products. You can't even find out the names of the coders working on it, much less contract them directly.
This is one of the most confused responses I know of on this subject. Not your fault, I hear it frequently. I would like to clarify some things for you. There are 2 basic types of OSS devs. Those that are employed by a company with some interest in OSS, and those that do it purely for fun. The devs in the first group don't develop for end users, they develop for their company. Somehow, through some process (in which they may be involved at some point) the Company decides what they have to work on, and when it should be done. This may have some benefit for the end user, it may not - and that is totally besides the point -- for all you know these devs are hacking away at some piece of code that will never be distributed outside of the Company anyway.
While this may be OSS development in the sense that people work on OSS code, it isn't about this topic - the "FLOSS Community" and the coders that form part of this community. Those coders tend to fall in the second camp. They tend to work for reasons other then direct cash. They do it for fun, peer recognition, whatever. For the majority of these people, "non-coding end-users" are the same bunch of clusterfucks they deal with everyday during the dayjob, and tend to not feature very prominently in the motivation chain. The things that drive them are project that are "fun" to code, "pet projects" and all that kind of stuff. They have little motivation to work on projects that are "boring", "seen as difficult" or "of no interest to the developers". This camp of OSS developers "must" do nothing, and more importantly, owe you nothing.
You then bring in some muddled argumentations about the "market" and "running out of business". Unless the OSS coder in question is pretty incompetent, and gets fired from his (quite possibly non-OSS related) dayjob, there is no "business" to be run out of. Most of these projects *are* pet projects, and they only reason you can use them for free is because the coder in question has an urge to tell the world: "Look what I can do!!"
Now its time to bring market drivers / basic economics into the picture. You, as a non-coding end-user, want an application. There are some half-way-there projects out there, but non really fit your bill. You are angry because all the selfish devs only think of their pet projects and having fun. Some entrepeneur, somewhere, will know this, and hire a bunch of devs to create a project you, and hopefully many others, will pay good money for. Only now, once renumeration has entered the picture, can you speak of a market in a meaningful way. Now you are a paying customer, and you can vote with your wallet and feet.
Unless you are a cheapskate, and don't want to pay for anything, but still want every little piece of functionality handed to you "just so". If you ain't paying the cash, either do it yourself, or STFU.
People who think they know everything are a great annoyance to those of us who do.
- Not-Invented-Here Syndrome: Reinvent and implement what already exists because it's not 100% the way you want it to be. Why collaborate when you can just create another duplicate project that will never make it past beta or even close to feature complete?
- Users, What Users?: Coding for yourself is nice, but if you want users to flock to your app, you might want to actually consider what they want. Don't bitch and moan at them when they offer suggestions, even if said suggestions don't fit your own personal vision, or even if they are downright stupid. That doesn't mean you have to implement them, but it means you have to be weigh them equally with your own ideas. Try to be inclusive and open to your userbase. "Go code it yourself" is a great way to keep OSS in the geek ghettos of the computing world.
- But It Looks Pretty: That's a snazzy looking interface you just whipped up, is it consistent? No? Does it follow standard UI principles? No? I'm sure people won't become frustrated and dismissive of your hard work. You can say that UI standards impinge on your freedom as a developer, but they make a user's life much easier, and makes people much more likely to actually use your software.
- Ask, Don't Beg: Asking companies and organizations to open code is nice and helpful, but be careful how you go about it. It can easily come across as "The OSS community could never dream of putting something like that together. Gimme!" Don't act like you *expect* the code, and that they are evil incarnate for withholding it. Don't make it seem as if the OSS community is incompetent and needs privately-developed projects turned over wholesale to get anything accomplished. Sure it helps a whole lot, but don't make it seem as though OSS is just mooching off the investment of others.
- Vendettas: If two projects can fight over something, no matter how petty, they will. Try coding, it's more productive and makes you appear like a mature, competent project that might help win over those hesitant to support OSS. Or you could just continue the pissing matches and flamefests over icons and licensing minutiae that could probably be settled if egos were set aside for a few moments. Public wars of words, endless forking....nothing gets accomplished but the stroking of egos. Well, except the whole "OSS developers come across as immature, childish amateurs" thing.
I don't think anything in it is wrong, as such, but it really doesn't say all that much. It sort of meanders through a few stories vaguely relating to the idea that "without an organizing vision, direction doesn't happen." But it seems to me that that while that's vaguely interesting, its not really a problem with the OSS community.
While, of course, the OSS community doesn't have a single vision for any piece of OSS software, quite a lot of OSS projects do and, as his story alludes, OSS projects that have a following but languish either for lack of vision or because the project owner has misguided vision—unlike closed-source projects which, while they may not tend to lack vision, are no less likely to have a misdirected vision than their open-source counterparts—can be rescued by forking.
And plenty of OSS projects do have a vision, direction, roadmap, etc. Sure, there's probably a lot of stuff that gets released under an open-source license (or straight into the public domain) because the author is essentially "done" with it and throwing it out to the community to do with what they will, but certainly open-source players like Apache, Mozilla, etc. have a vision for their main projects, and members of the community are attracted to and contribute to projects, no doubt, largely because of how they see the project's vision as compatible with their own. The "solution" McFarland offers is what it seems to me almost every major open-source project is already openly trying to do: allow the community to contribute, but institute a degree of top-down control in terms of timelines, roadmap, and assignments to make sure that the grunt-work necessary to have a polished project gets done.
I probably wouldn't call it "acting like the Cathedral", the openness of many successful projects to community process and innovation, while retaining a kind of top-down vision, is something of a synthesis: the Cathedral harnessing the energy of the Bazaar, the Bazaar borrowing the focus of the Cathedral. And you see something like it in the embrace by some commercial, formerly closed-source vendors of both open-source software and increasing community involvement. If I had to name the model, I'd call it the "Congregation" or "Assembly", a less-propietary Cathedral, a small portion of the Bazaar united by a common purpose and direction to accept, in the context of a project, some degree of authority and leadership (but not the exclusive ownership and control of the Cathedral.)
OSS doesn't suffer from the lack of leadership or other, supposed 'Cathedral' qualities. In fact, it's the superior leadership, based on merrit and ideals, that turns OSS into the nightmare of anything cathedral - such as MS.
In OSS much more than anywhere else, the best floats on top. That's why Outlook mail sucks and KMail sucks considerably less. Linux works because NOBODY doubts that Linus is the chief, Blender works because NOBODY doubts that Ton is the chief, because they both do an excellent job at what they do: leading large OSS projects.
Of course there's weedy stuff in OSS that's buggier and more twisted than Autodesk Converter and Macromedia Director together, but that sinks to the lowest bottom, and does not get pushed onto the market by monopolies and marketing budgets of galactic proportions (Windows XP anyone?).
The article is bogus and has it all backwards. I want my 5 minutes back.
We suffer more in our imagination than in reality. - Seneca
If you do this for fun then quit trying to foist it on everyone else. Quit trying to trick governments into developing and supporting OSS. Quit bitching about Microsoft. You guys got the commercial world interested now you have to live with the consequences.
Actually lack of leadership is a problem within the FOSS. Mark Shuttleworth has done a great job with Ubuntu because he is a good businessman. Too many FOSS projects are managed by developers who don't know how to manage which means poor decisions are made. This is fine if you don't want the project to grow however if you want to become bigger and better you need to make certain choices and sometimes they are not always easy to accept. I have seen many projects (both FOSS and commercial) die because of bad decisions being made by someone who has never managed anything in their life. Just because something is free and open source doesn't mean that they cant be managed by someone with a business background. Ubuntu is a fine example of this IMHO.
But with FOSS, I (the end user) can email the coder and offer to pay him/her to finish a feature I'd like or do some other boring job. And that is one of the great things about FOSS. Once I pay for it, everyone benefits from it (including me).
;-)
However progress will be slow because most of us will wait for someone else to pay for the changes we want. Most people will freeload if given the opportunity, Econ 101. Since you are reading this right now, I will thank you in advance for your future gifts to the community.
Because crossword puzzles have a solution.
It reminds me of how Canada, land of my birth, views itself and its relationship to the U.S.
It does somethings better, others worse. It will never admit what it does worse or will even defend it as a strength.
It's arrogant and sanctimonious even though it often has its heart in the right place. Other times it's naive in thinking that because it believes *its* way is right, it *must* be so.
It mocks the U.S. as backwards, even displaying a near pathological hatred for it, yet it secretly wishes it could hold the same lofty perch.
Now replace "It's/It" with "FOSS" and "U.S." with "Microsoft".
Frankly, what turns me off about the FOSS community in general is it reminds me of the science acolytes in the recent South Park episode when Cartman traveled to the future. Meet the new boss, same as the old boss. If Linux were king it would be Microsoft redux.
Feel free to mod me down, I have plenty of karma to burn.
You want to know who isn't running Firefox 2.x? They spell it "definately" and "rediculous".
FOSS is or is not better than closed source commercial code but the only way we'd ever know is to establish quantitative criteria and measure them with rigor.
Some things to quantitatively evaluate are:
Failures per release, time to release, bugs discovered, function points derived, cost-benefit, TCO, testability, verifiability, number of severity one bugs, number of severity one bugs never fixed, number of abandoned projects, time to next version, rate of customer abandonment.
There are probably 50 more I can't think of right now but the only sure way is to apply engineering and project management discipline to the criteria and comparison of those criteria. Then one must capture a candidate group of commercial and FOSS projects and track them over a multiyear period.
In other words we've been looking at the development experience instead of the results experience. How you build something is less important than what it does. Anyone who's ever seen the movie 'Apollo 13' understands this. More to the point though, development modalities reflect more the cultural aspects that the development team has almost no control over. Even in FOSS communities, they will self organize and operate according to features that have little to do with development.
We really don't know or care that much what the differences between good and mediocre closed source projects are. They are unverifiable in either case. So one cannot focus on the method. It's a black box. Instead we need to focus on the outputs and metrics that we can see.
1. Try Fedora Core, my mother, sisters, and best friend have all been converted, and all installed it themselves, and all have digital cameras, printers, and scanners that work with it.
2. Ease of use isn't everything. In fact, it's 1% of everything at best. People here act as though a learning curve is the end of the world. It is, if your goal is market domination. That is NOT an FOSS goal. FOSS cares not one bit about market domination. It cares about powerful tools that can be made to do complex work that its authors need done. In this, XP (or any Windows flavor) doesn't hold a candle to Linux (or to Mac OS X for that matter, or any Unix-family OS).
There is nothing wrong with Linux or FOSS. It's a brilliant success. The fact that some people are unable to take advantage of it does not make it any less so... and even that group is shrinking every day, thanks to the likes of Fedora Core, Ubuntu, etc. that are now as plug-and-play as XP ever was, and even moreso.
But that's an aside. The main point is: FOSS is more stable, more feature packed, more powerful, and free. Whether or not any one user as a test case is able to navigate it doesn't change these attributes.
STOP . AMERICA . NOW
Then there is Java -- long time criticized by RMS but not trumpeted by SUN as being GPL'd. And how about Open Office/StarOffice with a strong SUN contribution?
Yes, FOSS may be an insurgency, but I see it backed/encouraged/funded by big corporate entities who have a stake in its success.
One great thing about FOSS is that it tries to adhere to standards and have things be more or less portable. But people still use non-portable APIs (e.g. ALSA), and there are really not that many people with the knowledge and motivation to port things to Windows or Mac from Linux where most of the apps are written. I like GNOME, but saying that GTK is portable is only technically correct; you CAN port it, but it's not exactly easy or even possible in all cases. There is a Windows port, which is great, but the Mac OS X port is still not really usable (although there has been some great progress lately). I would love to see a fully ported GTK and QT for Windows and OS X, with established generic build systems; this would expose FOSS programs to a much larger user base, and allow more people to get involved in development too. These days it can be harder to set up the build system than to actually hack on the code. I have seen a few projects that have really focused on getting things to run on the Big Three platforms (Lilypond is my favorite example) and they have really healthy communities around them, partly because anyone with a computer can use the software.
That's exactly what I wanted to say.
I thought the article's comments about GNOME in particular were wrong. The problem with GNOME was not that it had no direction and therefore suddenly became bloated and unmanagable with feature creep. Far from it. It had a relatively popular leader who had an idea about how it should work. It gained feature creep because the "vision" of that OSS leader was to emulate a UI that itself was bloated, poorly designed, and suffering feature creep, both on the outside and the inside.
Had it had no direction, I believe it would have been more like the mix of UIs we saw in the early to mid-nineties on X11, or even throughout the late eighties on the Amiga. Someone would have put together a file manager. Another would have put an object viewer. Yet another might have worked on a print system. Each component would probably have been terrific, but the whole would have looked relatively ugly and argubly been poor in usability.
I'm surprised how "easy" it's been to fix GNOME and make it the relatively good system it is today (relatively as in recent versions, as configured by Debian and RedHat, are easily the second best mainstream GUIs, after Mac OS X.) That took the right kind of vision, and that vision, interestingly, was the result of the community noticing the project had gone badly wrong, and forcing itself to pay attention.
You are not alone. This is not normal. None of this is normal.
I thought Shuttleworth did a great job because he pays people money so he can tell them to do what he wants. Ya know, rather than just whining to people that they should have the same interests as you, he put his money where his mouth is.
How we know is more important than what we know.
I will state however, that a developer who don't know how to manage is still better than a manager who doesn't know how to code. Yes, in theory if you get the very, very best manager and the worst of all developers you could get a positive comparison but in any practical real world situation a developer still makes for a better manager than a simple manager, clueless managers are the root of all evil.
As for the SABDFL, Mark delegates all his actual managent task to his developers, he is just called in when the community is in a hippcup, but here he is much more of a client than a manager.
It should be mentioned that while a business person can organize a distro, the most succesful distros aren't run as business at all.
But... the future refused to change.
The bazaar model still worked when the pinnacle of software complexity were "cat" and "vi". That's it. It stopped working almost completely when complexity meant Open Office Org.
The Asperger's Syndrome kind of coder (and I'm one, so I can make fun of myself if I want to) which finds more joy in coding something cool instead of going out and flirting with a girl, also has a very narrow focus of attention and gets bored easily when he must deal with stuff either (A) outside that focus, or (B) which is basically homework instead of getting to the cool stuff. That's how we ended on the bad side of teachers in school, after all. Spending weeks understanding someone else's framework and code before you can even start on your cute "number paragraphs in Klingon" idea, is boring, and it's even more boring to understand and test all dependencies so you don't break something else.
So today in F/OSS the only ones making any progress nowadays are, sad to say, the Cathedrals.
Yes, everyone likes to use the Linux kernel and such as an example of why the Bazaar is strong, but have a look at the actual contributors some day. It's _not_ bored nerds like you and me working in their free time. Most of them are paid employees of Red Hat, IBM, etc. Linux as the work of bored nerds in their free time was a security shithole until Red Hat spent some real money doing a code and security review. And it was a joke in the enterprise arena until IBM started pumping some real money and formerly Cathedral-developped closed-source code into it. There's a reason why IBM looked like a believable target to SCO (as opposed to just a tempting target, by having deep pockets), and that's the sheer quantity of Aix code that IBM donated.
The same goes for OOo: practically all development is paid for by Sun, and it's bleeding Sun a ton of money. The same goes for Apache, which everyone uses as an example of why OSS is better than MS's software on a server: it, and most other Apache projects for that matter, is mostly IBM work. Go figure. IDE's? Both Eclipse and Netbeans are paid work by respectively IBM and Sun and a number of other corporate contributors. Compilers? You'd be surprised how much in GCC actually comes from Intel and the like. Browser? Mozilla was mostly paid work by Netscape, then AOL, and now it's mostly sponsored by Google. Etc.
So yes, as you aptly put it:
And that's why most of F/OSS nowadays is nothing more than a way for various corporate Cathedrals to pool their resources against MS. Sure, it's a good goal and I have nothing against benefitting from it. But let's stop pretending that ESR's Bazaar is anywhere _near_ relevant any more. The actual "Bazaar" projects are the thousands of unfinishet things on Source Forge that noone gives a damn about, either to help develop/debug or to use seriously or to pay the developper for features.
A polar bear is a cartesian bear after a coordinate transform.
Now, let me start out by saying I'm fully behind the F/OSS movement, and use it whenever I can. I'm working for a very large company, and we're using Linux boxes primarily. For that matter, I released a F/OSS package last year (WebFaceDB, available on SourceForge). But this is the first time I've been building and supporting as a straight sysadmin, not as a developer/sysadmin, and I've got to say that there's a *lot* of amateur, in the bad sense of the word, software out there.
/usr/kereros/include, since even the --includedir doesn't do that.
Let's starts with how much I *loathe* OpenLDAP, and the literally weeks I spent getting it working, and have yet to have it work with autofs, and I'm fighting it, right now, with Samba. I tried to find a GUI editor. The one that seemed best for my situation installed from an rpm... and had *no* useful sample configuration files, and even when I managed, using google, to set up some, it gave errors.
PHP4 (we have our reasons for not going to 5 yet) is a royal pain, and a *mess*. I mean, php.ini in */lib?!, and not in */etc? Why? Why scatter files all hither and yon? Oh, and then there's where I have to hand-edit the Makefile to add
On the other hand, Webmin was a literal no-brainer, and Nagios was only a bit harder. AND it came with working minimal configuration files. Even setting up virtual hosts with Apache were not *that* big a deal.
The amateurism covers things like inadequate testing, absolute requirements of a specific library (and not allowing a *later* version of the library), and not having an easy uninstall method.
It seems to me that a lot of folks push the envelope ->on their own system-, and don't try to meet standards that might run on nearly *everyone's* system. It doesn't have to be tested on *everything*, just follow standards. And to look at commonly-accepted practice, if not best practice.
mark, with more than two dozen years of software development experience, and half a
dozen with sysadmin