Slashdot Mirror
When Volunteer And Commercial Developers Don't Mesh
An "Anonymous" but easily identifiable
person submitted a commercial about a story at LinuxPlanet about
what happens when Volunteers and Commercial Developers have different goals over what is good for the Linux Community.unity and how it
might affect code. The article is largely about KDE and Corel. Please keep the flame turned down on this one, as it's a critical issue that still hasn't been solved.
Alright, I have several issues with this article, chief of these is the way the author goes on and on about not having beef with Corel then going out of his way to demonize them. Now ever since MSFT crushed Corel's Wordperfect with Word the Corel people have wanted revenge on MSFT one way or the other. Corel's current CEO is constantly criticized by business magazines for his foolish obsession with getting back at MSFT and their current involvement with Linux is simply part of this obsession.
Now Corel is paying designers and developers to work on turning KDE into a Windows-clone (i.e. a Windows beater), there's nothing wrong with that. What gets on my nerves about this article is the way the author trivializes the role of the KDE developers in this article. Corel does not pay any of the core developers salary nor does it contribute significant code that would die without their involvement. So basically the scenario is "Corel makes suggestions, KDE developers either like it or forward it to dev/null". If Corel really gets pissed off, they can fork the codebase (after all even Open Source developers have irreconcilable differences vis a vis Emacs and Xemacs) but if they realize they can't all they can do is keep making superficial changes and turning in bug reports.
Thus my question for the author is "So what's the big deal?". In any large project it is impossible for everyone to agree or have the same ideas and vision, simply because people have different goals doesn't suddenly mean some impending disaster will destroy the project. After all, IBM and the Apache developers have different goals but this hasn't stopped IBM from being one of the company's that gets it nor have their contributions been trivial. This article seems like a storm in a teacup to me.
Just my $0.02.
"To be honest, EVERY SINGLE PERSON who works on KDE is subverting it, in a sense. They are working on "...
The problem is one of centralized control. Corporations (and Entrepreneurs generally) seem to have an inherent need to control the process, and to make decisions that should be made on engineering (programming) grounds on other grounds. E.g., how does this fit with our product line-up. This tends to strengthen the corporation at the cost of weakening the product.
Good and bad are, of course, relative, and subject to decisions. I tend to consider it bad for a small group of people to be able to control a larger group of people. (Well, I also consider it undesireable for a large group to control a small group). The problem is weigh various trade-offs, and of course that gets quite complicated quite quickly (degree of control x number of controlled), and who is exerting what amount*kind of control is something where one can expect different individuals to have very different opinions.
I think we've pushed this "anyone can grow up to be president" thing too far.
YES, KDE developers are very grateful for Corel's input ! :-)
I can even prove it : I've digged up the best cvs comment about that: look at this one and this one.
The truth is not always well written in a nice article linked from Slashdot, it's sometimes hidden in CVS logs
The main trend I have noticed from this is that the number of RFC's produced has increased dramatically, and there are now competing RFCs and drafts sponsored by different commercial interests that are intended to handle basically the same problem. This is in part because some of the companies in question use the fact that there is an RFC for their implementation as a marketing tool.
The upside is of course there are a lot more resources devoted to working on problems. I would imagine that the situation with Open Source software will be better than that in the IETF because GPL is not (yet?) a sticker that can be put on something to make it more marketable.
Of course there will be problems. Corel is trying to make a product that is easy for the end user to understand. Like the line in the article about making their distribution look as much like Win95 as possible. There is nothing wrong with this. I do think that Corel should not be sending 100+ asthetics(sp?) issues to KDE.
I also don't think that there is anything wrong with Corel trying to "dumb down" their distribution. There are plenty of distributions that are fairly hard to install and use unless you've had unix experience. Some people don't care how the video card and the x server communicate, they just want it to work.
This happens within commercial companies too, BTW. What we do, when that happens, is move that person off that part of the project onto something else where he can't hurt the development process. I know that I find it irritating when some dude we hired off the street, who doesn't have any experience in our industry, starts telling the two seniormost programmers in the company "how it should be" without listening to a bit of our input as to what functionality the GUI needs. Unfortunately, the KDE team can't move this Corel guy to another part of the project, all they can basically do is "fire" him by putting him in their kill filters.
I don't know what the solution is, other than have people with experience in the free software community be the ones who the commercial companies hire to do the work. I don't think anybody has had any objections to the way further MTX development has taken place, for example, even though EST paid me to do the work of bringing MTX up to date to work with the latest/greatest tape libraries. On the other hand, people with experience in the Linux industry or with Open Source are rarely in the job market -- I know, we've been trying to find some of them to hire, and finding people who have the skills we need, in this industry, has been difficult.
-E
Send mail here if you want to reach me.
They've also contributed significantly to the WINE project, yet the author COMPLETELY missed that aspect. Hmmmmmmmmmm.....
Personally (and hold the flames, this is my opinion a concept many slashdotters have a hard time dealing with) I feel that WINE is a giant waste of resources. I don't think we should try to promote the user of the Win32 interface on top of Linux. Win32 is well known to have problems even running on it's native OS, why do we think we will make it better by putting Linux under it. WINE runs things just like they run on Windows. There is this persistant crashing problem that programs under WINE exibit (just like under Windows). Why do we want to turn Linux into a clone of Windows? I thought the whole purpose of Linux was to be an inexpensive alternative to proprietary Unix. If you don't think I know what I am talking about, feel free to email me (my real address is right there) and tell me what an idiot I am. I've heard it before and it still hasn't changed my opinion. Win32 should stay in Windows land. It was something I tried to leave behind, yet even under Linux applications are dragging it along.
I loved Corel when they had a real Unix/Linux version of WordPerfect, why did they have to use the Win32 version for WordPerfect 2000? Just another attempt to use Windows software under Linux, and it will make Linux look once again like a cheap (and bloated) Windows clone. Why?
Bite my yammer.
If you think you don't need functional abstraction, then the system you use has been so successful you're ignorant of its real complexity.
Linux could still be a lot easier to use, especially if it want to be on normal desktops. Microsoft, whatever its failings, has implemented a lot of good interface ideas in Windows. Using them will make it easier for users to switch away from Windows. (You know, embrace, extend...)
If a commercial venture like Corel can get Linux onto lots of desktops then it's a win for the open source movement. It may have to hijack KDE to do so, but would it really be a loss? Isn't the real goal of the KDE project to "dumb down" Linux anyway?
I bragged about my Karma at a job interview but I didn't get the job.
Ed Avis:
BTW anyone know how many copies of Corel Linux, WordPerfect and so on are being sold?
Point datum: The local Frys stores say that they can't keep Corel Office for Linux on the shelves, and the channel seems to be dry.
Lacking <sarcasm> tags,
Of course they are. This is nothing new; it's implicit in ESR's Bazaar model of open software development. The conflicts under discusssion right now are with corporate interests, but if every corporation on Earth went dark it wouldn't change anything because conflict is the heart and soul of creative collaboration.
The tug of war between conflicting interests is what makes for careful scrutiny, the "many eyes" that we're so proud of. Attempts to resolve conflicts are often the seed of brilliant innovations which make previous trade-offs moot.
By all means suspect Corel's UI suggestions, but keep in mind that many of them (such as deprecating inconsistent dialog button labels) are derived from very sound UI principles. We will need to provide Win-flavored UIs as learning aids, if nothing else; if this is anathema to the purists (perhaps including myself) then let's find a way to gracefully theme the problem away.
Peace and harmony are BOR-ING!
Lacking <sarcasm> tags,
This was bound to happen some time. There is an inherent conflict within every linux lover - they want to see it succeed commercially but they don't wan't it to go commercial. Till the linux community resolves this within itself (like by defining a core that is open source and let the rest be a commercial playground), there can be no real commercial progress that can satisfy both the communities.
Experience would suggest that anyone, whether corporate or individual, who gets involved with a project and makes suggestions; if that person/individual makes good suggestions and especially offer to implement them they'll be accepted. But if they are not particularly good suggestions they won't. Free software projects don't tend to accept patches if they are a hack just to solve some pressing corporate strategy.
Now Corel or whoever can go ahead and fork either the entire KDE or individual applications inside KDE. And the user community will either accept it or not. Most probably they won't because corporate maintained software doesn't tend to get maintained. So any sensible company should be desperately wanting to show a bit of diplomacy and get their changes integrated into the main project.
Of course there are bound to be conflicts when the commercial and open source worlds come into conflict. And it's going to happen a hell of a lot more as Linux becomes a more viable platform for corporate ventures. And, as a professional IT consultant, let me tell you that more and more companies, even some of the more staid ones, are becoming more and more interested in getting a piece of the open source pie.
The fact is that large corporations like what they see when they look at open source software - a lack of restrictive and expensive licenses, generally stable software and the kudos that comes from running this kind of solution. But what they don't like is the anarchic, socialist and often anti-corporate views held by the people that write and promote this software.
So their solution? "Collaboration" in open source projects, a method that allows them to "guide" the project in the direction they want - a marketable product. Unfortunately, whilst Linux hackers are quite willing to accept corporate backing when they need to pay the bills, at the slightest hint of a request for a change they immeadiately turn around and bite the hand that feeds them, as this whole KDE-Coral incident has shown.
In the long run it seems as though this dichotomy is going to force conflicts again and again, as commercial interests come up against the socialist ideologies of hackers. The only method I can see for this to change is for open source projects to incorporate themselves, allowing them to become their own masters.
If the KDE developers don't want to use the suggestions Corel gives it, they don't have to.
To be honest, I'd say the guy has a lot of valid points. And if he's been doing UI design for a while, I'd say he could probably bring a lot to KDE.
Contrary to popular belief, UIs are NOT easy to design so that they are intuitive for average users. The KDE team is working on it, I'm sure, but it can't hurt to have somebody help out. Christ, I'm sure there are some OSS projects that would DIE to have a professional UI guy helping out, especially one paid for by a major corporation. But instead of showing some gratitude, they write an article about how Corel might be subverting the KDE project.
To be honest, EVERY SINGLE PERSON who works on KDE is subverting it, in a sense. They are working on it because they want it to work in a way that will please THEM. Corel is no different than your average KDE developer. They want certain features, and they're going to try to get them done.
If they'd wanted to, they could've just started uploading changes to CVS, but they didn't. They filed reports about them, and gave the community the chance to discuss them. And to be honest, I agree with most of the ones listed in this article as being 'bad'. They make sense, especially considering Corel wants KDE to be ready for normal users.
Anyways, just thoughts.
-[Blaine]- "'Oh dear,' says God, 'I hadn't thought of that,' and promptly vanishes in a puff of logic."
Stallman says: "I think that corporate contribution to any particular free software development activity is always welcome, but corporate influence in impeding the extension of free software to do an additional job is obstructionism."
Raymond says: "I'm not worried. What I see is corporations realizing that if they want our results, they have to buy into our process--and if they don't, they'll be eaten by a competitor who does."
--- Hot Shot City is particularly good.
The objective of the distrobutions is not necessarily to produce comercial software (as part of the distrobution), but to give and identifying characterstic to the distro. For example, Corel's identifying characteristic is their file manager. (BTW, did the source ever show up for that?) Mandrake features DrakConf. Redhat pioneered kudzu in a distrobution (but it was also adopted by others). Everybody wants to have their distrobution be identifiably different from the others, and Corel's method is through customizing the file manager.
Corel's objective is not necessarily to produce a closed-source distrobution. They want to help the KDE project out, but in some respects they need to provide a different set of features to their customers than KDE does. Also don't forget that Corel's custom modifications to kfm are specific to how they have set up their distrobution in regard to samba support, where the filesystems get mounted, etc. They have a greater freedom to customize because they know how they have the system set up. KDE is distro-agnostic. Corel's desktop is a distro-specific hack of KDE. Is it their objective to fragment KDE or to destroy KDE? No, because they'd be biting the hand that feeds them. Remember that.
If anybody has a copy of Rhapsody for Intel to give away, drop me an email.
What particularly ticked me off in the article was the statement that because their UI designer was submitting all these bugs, one of the possible consequences for KDE would be a dumbing down. The users, it was claimed were good enough at determining what a good GUI was.
My question is : what user group is being talked about? Are we talking about the same user group that understands the X-protocol, and knows what to do with netstat, and thinks 'vi' is the best editor (it is, but there is no denying that it is hard to reach an intuitive level with 'vi'? These are not the same users that Corel is trying to win over. As long as the acknowledged goal of any software product is to increase its user base, I dont see any problems with a commercial company with the same goals filing such defects.
There is no such thing as luck. Luck is nothing but an absence of bad luck.
The early anecodote ("We've brought in a new person, no experience in what you do but he's a really hot developer, really, please bring him up to speed") has happened to me this past week, ironically enough.
I used to co-run a niche portal site, under linux, with heavy use of open source tools (phorum, some nice php stuff, some stuff I developed). The new owners brought in 'some hot programming talent' that turned out to be underpaid interns.
They also brought in a 'content manager' who was going to take over the site... but didn't know html and keeps reformatting the stuff using some windows program.
In short, we went from a friendly linux-backed site that required maybe an hour a day of simple script running to update... to a windows-2k driven monstrosity that requires a team of programmers and still can't update each week.
The lesson I learned? Folks with money aren't web-saavy, programmer-saavy, or programming-saavy. What they think of as 'hot talent' means someone who was good at schmoozing while others did the work. And since they fall for sales hype, they choose expensive, unworkable schemes instead of the simple, direct method.
If you sell something or let the VC come in, RUN! Disassociate yourself! If you stay, you'll just have regrets.
I must remain,
Anonymous
I am not qualified to assess whether Corel is a creeping black cancre on open source, BUT... being a profssional tester I am regrettably over-qualified to address the tone of hostility the author assumes towards testing. He writes in part:
"Changing file types is an advanced feature, and it should not be so easily accessible from a right-click menu," said one, on June 9. "A new user could easily handicap his or her system by accidentally playing around with settings without a clear understanding of their purpose." A bug? What here isn't working as designed? Or a dumbing down of KDE? Is there a single case of any KDE user ever "handicapping his or her system" by slapdash changing of file types?
It may be that the bug database for Debian is structured in such a way as to distinguish "bugs" from "design change requests". However, the supercilious attitude this developer assumes towards a voluntary bug submission is way too common in computing. Open source projects trawling for contributors often proclaim 'if you can't code, file bugs!' but if this is the attitude a bug filed in goodwill can expect to generate, don't be suprised to see your userbase ossify out to 99% hard-core developers and only 1% or 2% OSS newbies.
It's an unfortunate fact that development often looks down upon test, not the least because QA is staffed by typically less-educated or -skilled individuals. Keep in mind, however, that without this buffer of moderately knowledgeable testers between consumers and devs with their fingers stuffed in the code, many key issues of usability and quality will be pooh-poohed right out the door. The dialogs aren't consistent? Well who gives a fuck! They work, don't they? Ship it!
When you're getting QA input for free, don't look a gift horse in the mouth.
-konstant
Yes! We are all individuals! I'm not!
-konstant
Yes! We are all individuals! I'm not!