Debian Delayed by Disenchanted Developers

Update and modest suggestions by HighOrbit · 2006-12-20 07:51 · Score: 4, Insightful

There is an update on Andreas Barth's Blog that says "Update: There are media rumours floating around that "[Etch has] been delayed because some developers have deliberately slowed down their work". This doesn't reflect what I said."

The article did not say what packages were delayed specifically, but Debian is known to have an insane number of packages. Perhaps some culling is in order. I'm not part of the project, just an appreciative user, but here are my two cents.

Cut the distro down to what will fit on one CD (two max). That will reduce a lot of Debian's headaches. Less for them to maintain and less to test between releases. Everything else can be put into contributed non-official repositories
Don't be so anal and patch-happy with mainstream packages. Big projects like Gnome and KDE already do extensive testing upstream. Those packages should be able to move more quickly through the unstable-testing-stable cycle without sacrificing stability or extensive patching. How much of the debian patching on these type of big projects is *really* functionally necessary versus "I 'm the debian package mantainer and I want to put my mark on it".

About the project being "frozen", I don't know about that. I have a laptop running etch-testing. I did an apt-get dist-upgrade in mid-Nov , put it away for a few weeks and ran it again in early-Dec (don't remember exact dates). Something like 70 packages needed upgrades.

Re:Update and modest suggestions by nine-times · 2006-12-20 08:28 · Score: 4, Insightful

Cut the distro down to what will fit on one CD (two max). That will reduce a lot of Debian's headaches. Less for them to maintain and less to test between releases. Everything else can be put into contributed non-official repositories
I think that what's really great about Debian is that it has such wide support for everything. If there's a distro capable of being anything to anyone, and still doing everything pretty well, it's probably Debian. There are plenty of other projects that do just what you're talking about. They take Debian, reduce the number of packages to what makes sense for a particular purpose, and that allows more work to be done on fewer packages in less time, creating a distro that's more specialized. Why would you want Debian to do that, too?
Re:Update and modest suggestions by Hatta · 2006-12-20 08:48 · Score: 3, Insightful

Debian is known to have an insane number of packages. Perhaps some culling is in order.

You're crazy. The whole point of debian is so that you can apt-get everything in the freaking universe. I never have to go hunting down packages, that means a lot to me. Sure 90% of users only use 10% of the packages, but it's never the same 10%. So if you start just dropping packages, you're going to piss people off.

Cut the distro down to what will fit on one CD (two max). That will reduce a lot of Debian's headaches. Less for them to maintain and less to test between releases.

And less reason for anyone to use debian. If you want something that's pared down to a CD and doesn't offer you a lot of choice up front, try ubuntu.

--
Give me Classic Slashdot or give me death!
Re:Update and modest suggestions by arivanov · 2006-12-20 09:00 · Score: 4, Insightful

You have just described RedHat. No thanks.

I would rather have Debian release schedules, but have all the packages that are in it. Most of the sysadmins out there who deploy debian do it exactly because "Resistance is futile, you shall be packaged" and because "apt-get install light" works 99.99% of the time.

As a result there is a working platform on which to build services and commercial software regardless of what insane libraries your developers have chosen this time. Whatever it is, it can be apt-get installed. In the very rare cases you sometimes have to backport a version from testing, but someone has already solved most of the dependencies for you.

Trying something similar with RedHat quickly brings you into the land of RPM hell. I always love watching sysadmins suffering while trying to support development in a RedHat shop (especially where developers have su/sudo access). It is immensely entertaining to watch the network fall apart and be reduced to a random collection of machines all different from each other and each in its own circle of the RPM hell none being able to produce a release build.

So from the perspective of someone who has been running Debian driven networks for 6+ years and with 5+ years of supporting Debian as a base for commercial development I can say - no thank you, you misunderstood what brings most sysadmins to Debian. It is the best *nix development platform out there.

--
Baker's Law: Misery no longer loves company. Nowadays it insists on it
http://www.sigsegv.cx/
Re:Update and modest suggestions by srvivn21 · 2006-12-20 11:19 · Score: 4, Insightful

First off, I have nothing against Debian, and I don't advocate any changes to it's development model. I just can't abide baseless slander such as what you have posted.

You have just described RedHat. No thanks.

Yikes. This is so wrong. First, RHEL 4 comes on 4 CDs, not one or two. Second, many packages supplied by RH are patched so far that the original developers won't provide support on the mailing lists (Squid, OpenLDAP for concrete examples). Others are maintained by RedHat, which either makes them massively patched, or not patched at all. Neither of the points given really apply to RedHat.

I would rather have Debian release schedules, but have all the packages that are in it. Most of the sysadmins out there who deploy debian do it exactly because "Resistance is futile, you shall be packaged" and because "apt-get install light" works 99.99% of the time.

I'd bet that most of the sysadmins who prefer Debian do so because it's what they are familiar and comfortable with it...such as yourself.

As a result there is a working platform on which to build services and commercial software regardless of what insane libraries your developers have chosen this time. Whatever it is, it can be apt-get installed. In the very rare cases you sometimes have to backport a version from testing, but someone has already solved most of the dependencies for you.

Trying something similar with RedHat quickly brings you into the land of RPM hell. I always love watching sysadmins suffering while trying to support development in a RedHat shop (especially where developers have su/sudo access). It is immensely entertaining to watch the network fall apart and be reduced to a random collection of machines all different from each other and each in its own circle of the RPM hell none being able to produce a release build.

Am I to take it that you are saying Debian based systems are immune to this? Not so much the RPM hell (duh, Debian doesn't use RPMs), but the random collection of machines all different from each other even though the developers have root access? How, pray tell, do you manage that? Block access to the apt repositories?

So from the perspective of someone who has been running Debian driven networks for 6+ years and with 5+ years of supporting Debian as a base for commercial development I can say - no thank you, you misunderstood what brings most sysadmins to Debian. It is the best *nix development platform out there.

First, what does System Administration have to do with developing software? A Sysadmin's job is keeping the boxes running, not crafting applications to run on them. If a system admin WERE to develop software, perhaps he wouldn't use libraries that require such acrobatics his box is endangered? Second, big commercial software developers seem to disagree with you. For example, BEA, BMC Software, Hyperion, IBM, Sybase and Symantec, Lyris, VMWare, Oracle, and Elluminate. These are just software products that either I deal with on a regular basis or came up with in a quick search.

Why, if Debian is the best development platform in existance, would that be the case? Debian Stable changes at least as infrequently as RHEL, so it shouldn't be a matter of code stability.

Perhaps your dealings with RedHat based distributions have been less than plesant, but if you want commercial application support, it's either RH or SUSE. Tools for dealing with RPMs have advanced quite a bit in the last 5 years, and FWIW, I have no problems getting a bo

Heirarchy and human nature by heroine · 2006-12-20 07:57 · Score: 4, Insightful

Funny isn't it, how no matter how many times humans start over with a utopian system, they end up concentrating their wealth into a small number of strong leaders and leaving a large number of impoverished citizens. We really are programmed to institutionalize.

Re:Heirarchy and human nature by suv4x4 · 2006-12-20 08:21 · Score: 4, Insightful

If is funny. The question now is where do we go from here? Continue to be ashamed of our intrinsic natures and stick to faulty societal models (socialism) or accept ourselves as the selfish beings that we are and finally become comfortable with capitalism?

You see, both models are actually part of our intrinsic nature. As separate beings, capitalism makes sense. As cogs in a large system (or cells in an organism if you will), socialism makes sense.

Since we're currently on the borderline between separate beings, and part of one "uber-being" (society), such conflicts will always arise again and again.
Re:Heirarchy and human nature by the+phantom · 2006-12-20 08:41 · Score: 4, Insightful

Honestly, I question the sanity and/or sincerity of anyone who claims to believe that any pure system is "the answer," be it capitalism, socialism, communism, Christianity, Buddhism, atheism, or anything else.

--
Rhapsody in Numbers
Re:Heirarchy and human nature by poopdeville · 2006-12-20 09:28 · Score: 3, Insightful

There is no limit on how many times you can try again.

I was under the impression that most people are mortal.

Those who are on the bottom are those who constantly fail and hardly ever succeed or those who don't even try to begin with.

And with that, you completely discredit yourself.

To be a successful capitalist, you need *surprise* capital to start with. Not everyone is born with a silver spoon in their mouths. A lot of people work very hard and barely earn enough to feed and clothes their families. They certainly don't have enough to save for investment purposes.

What you're describing is called social darwinism. That has existed since well before capitalism. Perhaps you have heard of the French Revolution? It happened because *surprise* the French peasantry wanted to exercise their belief in social darwinism but were unable to, specifically because the Aristocracy held all the capital.

There wasn't a magic point in time when being a capitalist became legal or legitimate. It always has been, even in the Medieval period. Money talks, and everybody will listen. The French peasants simply had no opportunity to gain money, and thus comforts, without resorting to force.

For yet another example, consider American slavery. Indeed, it occurred during what's probably the most unregulated period of capitalism in America. Could the slaves try and try to be successful? Sure, but the social inequalities present during the period made such feats nearly impossible. But you had better believe that any rich black people in America were treated with respect for their money. Again, force had to be used to relieve the inequities.

So what is my point? First, you're obviously lucky. I don't mean to diminish your accomplishments, but it is clear (precisely because you accomplished them) that you were in a position to do so. Second, not everybody is as lucky as you.

--
After all, I am strangely colored.

Re:Dumb Editor by Cocoronixx · 2006-12-20 07:57 · Score: 5, Insightful

Understanding (or not) the behind the scenes nomenclature of a development environment has no bearing on your ability to use the final product.

--
"Obscenity is the crutch of the inarticulate motherfucker." - cloak42

Re:Pffft by croddy · 2006-12-20 08:01 · Score: 4, Insightful

I'm not even sure who's clamoring for Etch to release. Anyone who needs the latest toys can run it already, and anyone who really needs the stability of Debian Stable knows that it will be released when it's ready.

It's the other distros that seem to be in a huge hurry. To each his own; that's why we have more than one distro.

Dumb editor, but there is an issue. by Anonymous Coward · 2006-12-20 08:04 · Score: 5, Insightful

The problem is that dunk-tanc.org really is splitting the community. What they're providing is valuable to some - and does indeed help some problems - but unfortunately it's counterproductive to others people's needs and wants.

You've now got a subset of Debian guys motivated by money, and the rest of them still motivated by making a quality Linux distribution. Sometimes those interests are aligned (as the guys who set up dunc-tank observed) but sometimes those interests are NOT (as the guys who started Caldera and Novell now see when Microsoft can easily use the motivated-by-money lever to change the course of the projects).

IMHO, Debian should stay Debian - and stay as far away from money and paid work as possible -- and let organziations like Ubuntu build the corporate bureacracy stuff like release schedules, support contracts, etc. I hope Ubuntu buys dunc-tank.org and takes those employees with them -- because they and their work are useful for corporate marketing -- but do more harm than good to Debian development.

Re:Dumb editor, but there is an issue. by RLiegh · 2006-12-20 09:27 · Score: 5, Insightful

>So what you're saying is that Debian is for fucked-up smelly hippies who just can't handle the idea that people need money to live?

No. He's saying that he'd prefer that the people contributing to Debian are motivated by the desire to solve problems, and to make a good product better; as opposed to having debian be contributed by programmers whose attitude is "whatever, fuck it, it's good enough; where's my ten bucks?".

And he's not alone in that sentiment...not alone at all.
Re:Dumb editor, but there is an issue. by fm6 · 2006-12-20 10:59 · Score: 4, Insightful

That's one of the most self-righteous, idiotic statements I've ever heard. You're saying that anybody who gets paid to do something does it for the money and doesn't care about the quality of what they do. That's bullshit, of the smelliest variety. I get paid for most of what I do, but I take pride in my work. I've walked away from jobs — jobs were I was getting paid huge amounts of money — because there were other factors that made the job professionally or ethically unacceptable. And I'm not alone.
I'm guessing you've never had to worry about paying the bills or having a place to live. If you had, you'd know that sometimes people have to say, "God, I'd love to work on that, but I need to be doing something that brings in some money."
Re:Dumb editor, but there is an issue. by dubl-u · 2006-12-20 12:39 · Score: 4, Insightful

You're saying that anybody who gets paid to do something does it for the money and doesn't care about the quality of what they do.

There is a vast array of evidence that giving extrinsic rewards (like money) can reduce the quality and creativity of work when compared with intrinsic motivation. That's not to say that all people taking a paycheck will do shitty work. But I can list case after case in my professional life where I've seen reward schemes harm software projects.

For example, I recently charged some people a lot of money to clean up a mostly functional but hugely messy code base. The thing was almost impossible to debug, and completely impossible to improve. There were large amounts of what turned out to be dead code, a bunch of mismatched abstractions, and make-it-work hacks galore. What kind of idiot would build something like that?

It turned out that the programmer was perfectly smart, but the people who had hired him wanted the product really soon, so they structured it as a fixed-price deal with the price dropping every day. Naturally, he rushed, and by the time he pushed it over the line it was a terrible mess.
Re:Dumb editor, but there is an issue. by dubl-u · 2006-12-20 12:46 · Score: 4, Insightful

You're assuming that "bounty hunters", so to speak, would write worse code than people doing it for fun. That's an unproven assumption, and it's almost certainly wrong (as evidenced by the number of companies in the world writing software). Indeed, if someone stakes their livelihood on the quality of their code (i.e., they get paid for it), I'd say they're much more likely to be worried about quality.

I'm going to be charitable and assume that you have not actually spent a lot of time programming for a living.

As a consultant, I get to stick my nose in a lot of development shops, and I can pretty much guarantee that the number one cause of shitty software is people trying to do it on the cheap, by which I mean get the most apparent output for their dollars. New, quality-focused methods like Extreme Programming get a fair bit of their boost by making it much harder for the people with the checkbooks to exert time pressure on the programmers. (Instead, the time pressure is redirected into keeping scope as small as possible.)

That's not to say that open-source software is guaranteed to be of high quality. Some of it sucks. But you can be sure that you have removed a major cause of low quality, which is programmers giving up and saying with a sigh: "Well, that's not how I'd do it, but it's your money..."
Re:Dumb editor, but there is an issue. by Daniel+Phillips · 2006-12-20 17:40 · Score: 3, Insightful

Have you looked at the source of some Open Source projects lately? I cannot count how often I see error messages or assert failures with GNOME applications -- well, I start them on the command line and not in some window manager, so other folks don't seem to see them. KDE applications are not much better either; to wit the junk messages I just got from starting amarok:

QLayout "unnamed" added to QVBox "unnamed", which already has a layout
QLayout: Adding KToolBar/mainToolBar (child of QVBox/unnamed) to layout for PlaylistWindow/PlaylistWindow

First you talk about looking at source code then you show some assertion output to illustrate you point, which makes it sound as if you have never looked at any source code. Hardly gives you the credibility you need to go scoffing at anybody's software quality.

But let's consider your proposition: you seem to claim that assertion output (which you would only see if you run the app from a console) is some kind of indicator of software quality. Let me tell you what happens in a commercial development shop. The assertions never go into the code because such work is invisible and won't get the coder promoted. Who cares if some QVBox already has a message as long as the program doesn't crash when it runs? Answer: open sourcers do. Commercial developers tend to just want to bury the bad news and wait until it turns into bug reports. After all, closing bugs gets you promoted.

The assertion output you mistakenly characterize as proof of poor source code is actually part of the open source development process. Obviously, the first step in fixing a bug is to admit you have one.

--
Have you got your LWN subscription yet?
Re:Dumb editor, but there is an issue. by jschrod · 2006-12-20 21:09 · Score: 3, Insightful

First, I develop free software since 1983. Being a member of the LaTeX core team and being involved in development and maintenance of several high-profile TeX tools, I know what code quality is -- our code runs with very few errors, and obvious problems are resolved before release time, not after.
Second, the assertion is very fine if it is in. But that such obvious errors are not fixed before the release is rushed out is simply sloppy work that would not happen at our Open Source projects.
Third, as the CEO of a consulting company, I seem to work with better commercial developers than you do. And that is not only my own company, but also many other places where we work together with in-house developers. Yes, these folks put assertions in their code -- and they care for the case when they happen. That's because they are proud about their work and want to create good applications. They are not afraid of bad news, and closing bug tickets is not a metrics for their appraisal. (Client or user satisfaction is, actually.) Of course, there is a bad apple here and there; but that's not different to the authors of the thousands of OSS projects on SourceForge or elsewhere.
You seem to have a gripe with commercial software development -- you might have a bad experience in your own job. If you do, I have a recommendation for you: Look for a different company where developers are allowed to do their job. It pays in the end, both for the developer and the company.

--
Joachim

People don't write Manifestos any more -- what's going on in this world? [Frank Zappa]
Re:Dumb editor, but there is an issue. by Ed+Avis · 2006-12-20 22:51 · Score: 3, Insightful

The FSF has employed paid developers for years. It doesn't seem to have distracted them from their goals of free software or led to them being corrupted by Microsoft.

--
-- Ed Avis ed@membled.com

Re:You work for free, or... by drinkypoo · 2006-12-20 08:39 · Score: 5, Insightful

"Many unpaid developers simply put off Debian work to work on something else."
Please, correct me if I'm wrong... but isn't the whole point of Open Source to contribute code for the betterment of the community? Which, as it happens, means not getting paid to write code.

Open Source is a development methodology. Free Software is a moral standpoint. Neither one says that you can't get paid. Neither one, in fact, says that you must do anything for the betterment of the community - once the appropriate license is used, EVERYTHING you do with the program that is legal contributes to the betterment of the community.

In fact what you and many other people miss is that no one does something for nothing. Sometimes they do it just because they are addicted to the good feeling that they get when they do something altruistic, but at the base level, they are feeding a stimulus-response pattern in their brain that causes them to want to do that. They are being paid in good feelings.

If I am contributing work for which many people get paid, and then I see that someone else is being paid for work which many others contribute, I may come to the realization that I need to pay my bills and they cannot be paid with good feelings which are unfortunately non-transferable and not considered legal tender for any but the most private of debts, if you know what I mean. Or maybe I'll just turn into a stingy bitch who wants some of that or y'all can fuck off. Either way, the contributions don't get made.

Ultimately, if you're going to have a release schedule and you plan to stick to it, you're going to either have to pay some people, or make sure some people don't need to get paid, which boils down to supporting those people, which is a form of pay even if you don't give them actual money. Otherwise you will have problems because people will have other motivations. This will continue until the cost of living drops so far through technology that people no longer have to work. Then we will have new problems.

--
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"

Re:You work for free, or... by davek · 2006-12-20 08:42 · Score: 4, Insightful

Can complex software really be done in your spare time? That really is the question, isn't it? If the answer is "no," then it seems like open source software is what the critics say it is: an anomaly created by the birth of the internet, and it will die out like any other fad; leaving established, commercial software as the primary source of usable software technology.

Let me be crystal clear: THIS IS NOT TRUE!!

What is happening is the value of software is shifting. In the future, you won't have to work on open source software "in your spare time." You will be paid to work on open source software by the company you work for, because they have a stake in the software's success. Software is a living thing and must be maintained. If my business directly depends on... say... Asterisk running correctly, then I'd better have at least one OSS hacker who knows the Asterisk source code... get it?

Remember the old mantra: Free Software was never intended to be free-as-in-beer. You still have to pay for it if you want any real commercial use out of it. Companies will slowly realize they don't have to pay a monopolistic empire for all their software needs, but rather can hire their local blue-collar OSS hacker. Only then will the economy make some progress...

-dave

--
6th Street Radio @ddombrowsky

Re:Dumb Editor by walt-sjc · 2006-12-20 09:14 · Score: 4, Insightful

This is why Linux will never catch on. "Packages going into Etch"?? WTF does that mean?

Genuine Advantage in Vista? WTF does that mean? This is why Windows will never catch on.

iSight on a Mac? WTF is that?

If a well-educated slashdot reader has no clue what you're talking about, how is the general public, let alone my grandma, supposed to use Linux?

I would bet that most Linux using and a large portion of non-Linux using slashdot readers knew exactly what that meant. By your trollish and poorly thought-out comment, I would assume that you are not in the majority here. Terminology in technology always requires some domain knowledge. This article is NOT aimed at your grandma (doubt your grandma reads /., "news for nerds",) and would have no bearing on her use or non-use of Linux. It is an article about internal politics of a particular distribution of Linux that she probably wouldn't be using anyway.

Re:Dumb Editor by Fnkmaster · 2006-12-20 10:10 · Score: 3, Insightful

If you want the latest, greatest "Vista" from Microsoft, then you do need to worry when they keep delaying it. Likewise for Debian - if you want the latest, greatest "Etch" release, you need to worry about delays too.

If you don't care about that stuff, then just use the version that's available right now.

I fail to see how this is any different between MS and Linux? With Linux the process is more transparent, so when you go to a site that bills itself as "news for nerds", you can read about all the gory details if you want.

Slashdot Mirror

Debian Delayed by Disenchanted Developers

23 of 329 comments (clear)