I saw this video a while back, and to be honest, it made me avoid subversion.

I have gone from CVS to git. They just seemed a bit too oiley and arrogant over the whole matter. That guy could have forked subversion had he really cared, and a date parser is a real minor issue to make this mountain of mole hill over.

Truth be told most open source projects are written by small teams of people, some seem to like making up their own rules to fit in with what makes a good project for them, it happens all over. It is a small minded to apply the term poisonous to anyone disagreeing, and then going off lecturing about it really.

Subversion, heh, just another versioning system to have installed to get at source code, git stole it's thunder a while back, and CVS was workable at Subversion's zenith, I think they may just be a bitter over it all.

I've worked with many a menace to productivity in my days, people who will spend much more time arguing about how the job should be done and trying to talk you around to their way of thinking than it actually takes to code the damn project !

To me, what this actually sounds like is that the boys at the top want total control of the project, and only want their names associated with it.

That way they get all the credit/benefits, for example cosy jobs with Google.

Seriously. Look at what they said.

They'll accept no discussion - the only valid ideas/comments they seem to want are the ones *they* put on a site.

And of course no-one can have their name on anything - except of course the 'founders', who seem to have the idea that they own everything while simultaneously claiming that it belongs to the community.

I guess making comments, getting credit and acquiring personal benefits on an open source project is OK, but only if you're the right people.

Personally I think they can go f*ck themselves.

I've worked on a few quite unstructured open source projects. Largest only consisting of 6 developers. However, it was a nightmare to sort out how things should get done.

Funny part is when people go tagging poorly written code. One line in particular caused me hours worth of trouble, hidden in a little file causing strange behaviour because basically someone wrote a line to reset a variable with the variables the wrong way around. Took 6 hours to find that, then I found it tagged with one of the few dev people that heavily tag they're code.

Some people are literally hostile when it comes to defending their tags... yet they seem to lack in documentation.

On the video, they mention lots of cases of poisonous people but they don't give names. However, they also mention that all the mail is in the archive. So who are the poisonous people, and how poisonous are they in practice? I think it would be interesting to review the cases.

By the way, Subversion is pretty useful, much better than CVS or RCS. I don't know how it compares to Git, SVK, or other, more modern, systems, but the big advantage of Subversion is that the bugs have been taken out. I also recommend "TortoiseSVN" on Windows - very handy, very easy to use. The documentation is free too.

I look at it from the developer's side. All they sometimes get from all the hard work is recognition. Who am I to deny that? They have to eat too, and I actually prefer myself to employ people who can prove they've coded in Open Source because it means they are capable of collaboration and integration (and I'll figure out the personality issues myself, thanks). The approach as described in the article would deprive these people of the one benefit (other than experience) such project could bring them.

This comes, of course, with a caveat: each piece of code must be seen as a contribution, not as a territory. It's courtesy to debate a change to someone else's code, but "ownership" must not get in the way of improvements, and THAT arbitration is the true role for a project leader - the big picture.

So, philosophically I can understand what they try to do (although it looks rather self serving from one angle) but I don't agree. And that's fine. Differences of opinion make for debate (not arguments, debate) and that is just as critical for quality as good coders. There is no better way to kill the quality of a deliverable than to stop debate or prevent dissent from being listened to.

Just look at the first days of Terminal 5..

Who died and made these two the leading authority on open source development?

We shall continue to develop software the way we always have, stop telling us how to do our jobs.

There are NO absolutes, but I don't see a problem with letting your contribs add their name to their respective modules, and to suggest that they'll become possessive over each one, is simply wrong.

This would seem to explain why many open source projects never quite feel finished. There are always a few silly little faults and niggles left un-squashed. Usually it's nothing that actually undermines the functionality, you understand, but it all adds to the impression that the software is only held together by metaphorical bits of chewing gum, sticky-tape and string.

After all, if you're forbidden from attaching your name to your work, why bother spending time tidying it up beyond the bare minimum?

[...] believe in open source projects that maintain a small "bus factor." That would be the number of contributors who could be hit by a bus before a project collapses. [...]

so they want the project to collapse if just one contributor is hit by a bus, in contrast to only collapsing if most contributors take the bus? somehow the definition got reversed i think. the bus factor should be defined as 1/redundancy or 1/(number of brains every piece of knowledge is contained in) (this would be a wetware RAID). or they should aim for a big bus factor.

Well I've used Subversion for quite a while and CVS too for many years. Subversion is one of those projects that you would never want to work on - if you screw up, people lose their source code and that's _bad_! Like an accounting system:-)

Subversion sticks everything in an opaque database that is an island all to itself and only provides one real improvement over CVS - the clean handling of binary files.

The win over CVS is marginal, certainly no order of magnitude advantage which makes one wonder why these guys think they're so hot?

Authors claiming credit for their work has a long history and gives the author a reason to make sure what they write is good. If it's flawed then it's OK, you can hide under the anonymity.

Perhaps the Subversion guys should try to develop something truly innovative rather than just another revision control system. The commercial ones are leagues ahead anyway.

Oh sorry, I forgot. They landed jobs at that flavor of the month advertising company - Google, so all they say must be taken as true. What software innovations have come out of that place again? Oh yeah, how to sell ads, how silly of me.

That these _are_ fairly poisonous idiots intent on plastering _their_ names

on other people's work. Hairy T once said "its amazing how much you can

get done if you don't care who gets the credit". Something these blokes

appear to have missed. And Subversion? I think I'll keep my eyes open

for something better and a little less anal.

~D

Can we not fork subversion and run a branch that DOES encourage ownership of code and pride in your work?

Thing is about open source, you can also protect it from the people who think they control it....

Lots of FLOSS projects see the "poisonous people" problem as they grow, and if you're not careful in controlling it then you risk your project being paralysed. Poisonous people are the ones who can suck all the life and fun out of a project, both wasting the time of useful existing developers in pointless discussions and discouraging new developers who don't know up-front which are the people they should be listening to / working with.

Subversion has only one improvement over CVS? You've clearly never used these for projects of any size. Subversion has transactions - critical for any real system with many developers. It's orders of magnitudes faster to update on large projects. Branches are extremely lightweight and fast to create. Directories are first class structures, so you can actually move and delete them. Etc...

I'd just about started using it when I got pushed at 'git' because of the kernel. Its quite an amazing piece of code; I don't know it that well but I think it may turn out to be an object lesson on how Open Source projects should be done.

Google will be less important one day and their geeks will want another job. Let the names forever be known.

Just what I was thinking. Ideally, an arbitrarily large number of people could be lost without affecting the project, so you want a high bus factor.

They do have a point that possessiveness of code can be a serious problem. I think it would work well to ban attributions in the source code, and also disallow anonymous updates. You have accountability, e.g. the ability to look up who's been submitting poor-quality code. If you want to show off what you've done, you can find all the updates you've made. And it's not easy to identify who wrote a particular piece of code, so claiming 'ownership' of it is much more difficult.

Yes but are they on the bus or being run over by it?

(I think there's still erroneous references to small bus factor when you mean large)

Having people's contributions recognised (Developed by blah with help from blah and blah and blah; Thanks to blah for blah) is one thing, and is Good. Even if the code someone used has been replaced entirely, they still contributed, and deserve recognition.

Tagging authorship directly in sections of code is quite another. For a start, people will start emailing those people directly instead of using exisitng channels (which can mean potentially policy-affecting code gets written without being discussed with the maintainers / head devs and/or the dev community at large. Secondly is the "my code!" aspect, where anyone touching the date parser gets flamed by the author because that part's got his name on it. Thirdly is the situation where the code DOES change a lot but still has this person's name on it, meaning the attribution is no longer correct (and ties back into the directly-emailing-instead-of-addressing-the-dev-community point)

“A simple means of maintaining a small bus factor, they say, is banning names from source code.”

That looks suspiciously like a second instance of that mistake.

(For some reason, I'm reminded of one of Terry Gilliam's animations, in which a bus was tripped up…)

We are in charge, anyone who disagrees with our notions of how things ought to be, no-matter how right us will be labeled as "poisonous malcontents" in front of the group, since the other developers are a bunch of sheep, they will go along with it based on the weight of our authority. Sounds just like a real job. Just another group of folks looking to hold onto power at any price, and whose cannot possibly measure up to the rules they have set for others. Why in the world would I want to subject myself to that for free?

"Can we not fork subversion and run a branch that DOES encourage ownership of code and pride in your work?

Thing is about open source, you can also protect it from the people who think they control it..." ..... By Anonymous Coward Posted Saturday 31st May 2008 16:25 GMT

That would be Underground Counter Insurgency Work, AC, and you would be Liable and Fully Entitled to Unlimited Funds/An Open Ended Line of Credit for All Needs and Feeds. The SMARTer Ones Topside would then be able to Do their Thing and Magically Arrange Fantastic Dashing Flash Cash and we get to see their Hand and any Grip they would Hold on the Concept.

Well.. The question is..

What if these two @sses get hit by a bus? Who do we turn to? NO IDEA because there is no mention of other names anywhere because names are just so damn poisonous!

This is just wrong.

I always think its nice to see the personal side of software. Just because you have a name it doesn't mean your a d!ck.

If I was working on some code and saw that Joe Bloggs had written some awesome code, a email his way saying "wow this code looks great" could really make him feel good about HIS PERSONAL contribution to the community/project.

Respect the people more, it doesnt mean they want to destroy the project because they would like a mention somewhere. One name in commnuity projects like SVN is just a drop in a pond.

"In Subversion we have no name. Elsewhere we have a name.. My name is Rob Cooper." (Fight Club-esque!)

I don't understand why people are so outraged about this idea of someone "owning" an open source project. Why shouldn't their word be law?

They's dead right about "attention and focus". As nice as it would be to have a perfectly consensus-led project, the gains of having someone competent (assuming they are competent) keeping everyone working towards the same goal are significant.

If you don't like it, fork off ;)

every time he or she commits a change.

So if you have the repository it is hardly difficult to find out who touched a particular line of code. It can be tedious at times, but not difficult.

And Subversion in fact makes this job easier than most, see: http://tortoisesvn.tigris.org/blame.html, for instance.

So the whole discussion is about a complete non-issue. Anyone who wants to know who did what can easily find out.

I use bzr.

<quote>

There are NO absolutes, but I don't see a problem with letting your contribs add their name to their respective modules, and to suggest that they'll become possessive over each one, is simply wrong.

</quote>

Pardon me sir, but your philosophical slip is showing! :-)

The tweed jacket with the meershaum pipe in the pocket, please.

"There are NO absolutes" ... that's right, absolutely NO absolutes - absolutely stated :)

I'm all for what these guys are doing, whether adding names to source is right or wrong, I don't really care, point is they've found what they think is an indicator of poisonous people. Find it, take it out, move on. If you disagree with their policy, then here's a hint - the test is actually _how_ you disagree, and looking at these comments is quite an education in that respect.

There's a lot of projects which have to deal with poisonous people, and gentoo is one where a lot of damage has been caused by them.

There's definitely a balancing act to keep talent without bogging down on politics, but having seen the results of a lot of people's unchecked vitriol, I think I'm glad that at least some people are finding ways to keep a tab on these things.

The whole purpose of forming one of these open-source communes is to assure that no corporate bourgeois overlord-oppressors dictate to the workers. By being an open-source commune, the spurned workers can at any time take full ownership of the means of production as a fork of the open-source commune's source code.

It sounds as though Brian Fitzpatrick, Ben Collins-Sussman, and Google need to be taught a lesson that they are not permitted to be corporate, bourgeois, overlord-oppressors. Subversion programmers of the world unite! Rise up against your oppressors. Take control of the source code of the commune! Fork Subversion. (And, you new Subversives-no-more, please rename your new fork to some more-appropriate name for the business world, such as Terrorism or Racism or Perversion. Seriously, perhaps Interversion or Introversion.)

I've never contributed to open source, but for the last 20 years have been tagging code with my initials.

why? in the team if someone didn't understand a piece of code they could have a chat with the guy who wrote it (or express displeasure with the way it was written, depending on how long the debugging had taken!).

Now with Subversion or similar there's less need for such tagging, but it's still convenient.

MAH

Let's see, a couple of folks decide to start a FOSS project and put their vision into action. They decide how it will be run and in what direction using whatever roadmap and contribution rules. Others want to contribute without forking the code, so they should a) Rant and rave and create issues within that developer community until they either get their way or are forced to leave, or b) Follow the original rules of the project and contribute to the Bigger Picture (tm)!

On the flipside, I totally agree it seems anti-FOSS to not let folks take ownership of their contribution AS LONG AS they are not viciously possessive of it. Folks are going to critique it, are going to change it, are going to mangle it, are going to learn from it (whether it's how or not how to code), and whatnot. In a meritocracy where your code says how much you can contribute, I'd want to know whose neck to strangle when some snippet or module makes the whole thing crash in flames of un-glory.

"On the flipside, I totally agree it seems anti-FOSS to not let folks take ownership of their contribution AS LONG AS they are not viciously possessive of it." .... By Tom King Posted Monday 2nd June 2008 15:33 GMT

Tom,

Surely by automatic Default of Contributing to FOSS is that vicious worry unfounded and experienced coders will be well aware of the Honey Trap which Tempts Self before Service... and they will Attend to Business First and then Sample the Improvement in Its Pleasures.

"But he insisted on tagging the source code with his John Hancock."

Not a coder so may be completely off track, but:

Surely if someone is going to get all prissy about "their" code they will do so whether it has their name on it or not? Surely a coder can recognise something that they have written and see any changes that have been made from their submission.

Also, does this whole premise not diminish the appeal of FOSS with respect to proprietary software? The 2 main appeals of FOSS are the fact you can see the source code and the fact that the community is involved in the creation process.

If you take the community out and revert to a "my way or the highway" approach then you are very likely to alienate a lot of the best thinkers and coders; and even if good thinkers stay if you don't listen to them then what is the point? It is like proprietary software without the benefits of being proprietary.

IMHO, obviously

already pass a copy of the communist manifesto around?