This is a weblog I'm keeping about my work on Debian and any other useful Debian related info I come across. It is not meant to compete with other news sources like Debian Weekly News or Debian Planet. Mostly it is just a way for me to classify and remember all the random bits of information that I have floating around me. I thought maybe by using a blog it could be of some use to others too. Btw. "I" refers to Jaldhar H. Vyas, Debian developer for over 5 years. If you want to know more about me, my home page is here.
The name? Debain is a very common misspelling of Debian and la salle de bains means bathroom in French.
If you have a comment to make on something you read here, feel free to write to me at firstname.lastname@example.org.
You can get an rss 0.91 feed of any page in the
blog by appending
?flav=rss to the end of the URL.
If you want to do as Joey Hess did and import Advogato entries into your blog, you should take his advice and use the XML-RPC interface. To do so, may I suggest my WebService::Advogato perl module? I haven't gotten around to getting it in Debian yet, though the source does include a debian directory. A simple script to include all entries might look like this:
my $client = WebService::Advogato->new('user'. 'password');
my $numentries = $client->len('user');
if ($numentries > 0)
for my $index (0 .. $numentries - 1)
my $entry = $client->get('user', $index);
my ($date_created, $date_lastupdated) = $client->getDates('user', $index);
# Now munge into whatever format your blog uses.
Of course you should replace 'user' and 'password' with your real username and password.
The big news in Debian of late has been the meeting of members of the ftpmaster and release teams in Vancouver and the proposal for post-sarge releases that came out of it. This spawned a response of over 1000 messages which I've more or less ignored uptil now if for no other reason than such a monster kills IMAP threading with Dovecot. (Admittedly my mail server is rather underpowered which is a big factor.)
Well, last night I arrived in Atlanta and I was bored and couldn't sleep and my hotel had Internet access so I read the Nybbles thread. And read and read.
I have to admit that like many, my first reaction was consternation that we are "dropping architectures." I felt the same way initially as I have just gotten through telling lots of people at Linuxworld that one of the reasons Debian was so much better than Red Hat, is that we don't just drop arcitectures. But if you read the announcement and the clarifications posted afterwards, whatever dropping is to be done is possibly only temporary and if the porting teams can get their acts together, Etch may have just as many architectures as sarge. There seems to be some concern that the inclusion requirements are arbitrary or set too high but these are implementation details and it is after all just a proposal, not a fiat as some people seem to think. It will all be worked out in due course.
Why did it come to this? Unfortunately, for all our vaunted size, we simply do not have the manpower to keep all these architectures in sync and release in a timely manner. (Most Debian developers only care about their own packages. Relatively few work on "big picture" issues.) This proposal is an admission that for now we have bitten of more than we can chew and from now on growth must occur in a more restrained and well-planned manner. There is no cause for gloom and doom. We are still growing. The Debian product will not change in drastic ways.
As you can see, I now feel that the Vancouver proposal is a good thing minus a few blunders. (I mean, "Second Class Citizens"? Nobody at the meeting realized this would be a red rag to bulls?)
It's also time to think of "second class" maintainers. Many of the developers are simply not capable of/not interested in/don't have the time to do the hard work needed to convert a big blob of software into a stable, well-integrated operating system. Yet they have the same weight as people who work their arses off for the project. Tier 1 developers should get more powers to do NMUs and other things necessary to do their jobs without bureaucratic hassles. They would need to renew their status in some way at say 6 months to weed out deadbeats. Conversely Tier 2 developers would be easily be able to step up to Tier 1 if they wanted. These labels would cast no aspersions on the abilities or characters of the labelled. They would just be an indication of the amount of investment the person is able to make in the project. I have no illusions anything like this is going to be adopted anytime soon but it's something to think about.
The other big political thing going on is the 2005 DPL elections. Here's how I'm going to vote in reverse order.
(7) Jonathan Walther
He manages to alienate everyone he comes across. Not a good quality in a DPL.
(6) None of the Above.
(5) Angus Lees
Not understanding the constitutional process of nomination and failing to provide an initial platform wasn't encouraging.
(4) Anthony Towns
Would make a great DPL but a large part of his platform involved being nice and people otherwise being trusted to do the right thing and sadly that just won't fly in the Debian we have today.
(3) Matthew Garrett
His platform intrigued me. He has a common-sense approach to the DFSG which I like. But he kind of wussed out in the debate and gave safe "politician" answers like everyone else.
(2) Andreas Schuldei
Also nice. Next! What do I have against nice people? Nothing at all in real life but Debian is a huge, sprawling, strongly opinionated bunch of people. It takes a forceful personality to sort out that kind of mess. Anyway Andreas and Branden are campaigning as a team so he may yet be in a position of influence.
(1) Branden Robinson
Branden has the qualities needed in a DPL. He is passionate about our ideals, able to express them articulately, and charismatically both within the project and to outsiders. He takes the initiative to get things done. In particular with the whole SPI treasurer debacle he did the best he could with a bad situation. Branden gets my vote.
This year Linuxworld is moving from New York to Boston. If you are local to the Boston area and you would like to help out at the booth, let me know and join the email@example.com mailing list.
If you would like to help maintain the Debian webmin and usermin packages which I orphaned a couple of month ago, get an account on alioth and let me know so I can add you to the pkg-webmin project. You don't have to be a developer but a good knowledge of perl will help as that is what this software is written in. A mailing list has also been set up for coordinating maintenence.
So I came home from Connecticut and opened up debian-devel for the first time in four days to find a General Resolution had been proposed to get the amd64 port into the archive and a huge flamewar has resulted. The amd64 porters feel they are being blocked by ftpmaster intransigence and lack of communication. Other people are sickened that Debian, which has always prided itself on doing the right thing not necessarily the popular thing, is now using voting as a means of coercion and the ftpmasters are unfairly being hassled. Who is is the real victim?
I'll tell you who the victim is, it's ME! Why am I forced to have to sift through large piles of crap, just to keep abreast of what's going ? Just kill the threads you aren't interested in I hear some people say. The problem is threads drift. Occasionally, there is some useful and important bit of information amongst the drivel. (Again, remember ajs' social contract bombshell? Delivered during a massive and unrelated tread on debian-vote.) I am subscribed to 18 mailing lists and hang out on IRC a couple of times a week and I still don't know exactly what the deal is with things like the amd64 port. This is a major development one should know about don't you think? If I could devote my life to the project as some people seem to be able to it wouldn't be a problem but I can't do that. God, family, rent, Debian has to take a back seat to all of these nowadays. I don't like this situation. Working on Debian has been both pleasurable and brought me material benefits but something will soon have to give. Of course I could just follow the lead of atleast 50% of the developers (according to recent voting patterns) and just live on the little island of my own packages not paying any attention to what goes on within the project. But doesn't such a view make a mockery of our ideals of openness and collaboration?
Manoj says we should have a little sympathy and respect for the people behind the titles. This is sage advice and I agree. But how about if those people showed a little sympathy for us? I'm sure writing status reports is a boring waste of time but think of how much time is wasted by other people when they aren't written. For a great example of something I wan't to see more of, read Roger Leighs' report about the LSM Free Software Printing Summit. (The only criticism is that it should have been posted to debian-devel-announce.) Frankly if people in key positions can't do something like that once a month, they are a hindrance and should step aside or be replaced. Debian is not a little club anymore. The amount of communication which was adequate in those days is not adequate anymore.
Today is the last day of voting on the General Resoluctions concerning Sarge. Here are my choices:
-=-=-=-=-=- Don't Delete Anything Between These Lines =-=-=-=-=-=-=-=-
[ 2 ] Choice 1: Postpone changes until September 2004 [needs 3:1]
[ 1 ] Choice 2: Postpone changes until Sarge releases [needs 3:1]
[ 6 ] Choice 3: Add apology to Social Contract [needs 3:1]
[ 3 ] Choice 4: Revert to old wording of SC [needs 3:1]
[ 5 ] Choice 5: "Transition Guide" foundation document [needs 3:1]
[ 4 ] Choice 6: Reaffirm the current SC [needs 1:1]
[ 7 ] Choice 7: Further discussion
-=-=-=-=-=- Don't Delete Anything Between These Lines =-=-=-=-=-=-=-=-
The reason is that we have get Sarge out right now. Sorry the rotting corpse of woody is beginning to smell. Once that is out of the way, people can do all the legal noodling about they want. I prefered choice 2 over choice 1 because in the event of something going wrong a vague date is more flexible than an exact one.
My next preference was choice 4. Even though the changes were strictly editorial in my opinion, the release manager doesn't agree. And only chose to mention it at the local planning department on Alpha Centauri ^W^W^W^W^W^W^W^W in the middle of an obscure thread on debian-vote. So if it's going to hold up sarge, let's get rid of it.
My next preference is choice 6 which is to let the editorial changes stand. I would want to force the release manager to change his mind or step aside in that case because as you may have gathered by now, sarge must be released right now and it is way too late to do drastic surgery on it.
A transition guide is just extra bureaucracy and an apology is just plain weaselish. So no love for choices 5 and 3. And last of all choice 7, further discussion because the last thing we need is even more gum-flapping on this topic.
Bah I say! Bah! A vote on the document that defines the very foundation of Debian philosophy is met with resounding apathy. Amongst the ones who did vote some are saying "But we didn't know we were voting for _this_." (Come to Jersey City. The local Democratic party machine loves people like you.) And then aj has to go and posit a totally far out interpretation of the (modest--I actually read them) changes to the social contract that might require us to delay the release of sarge until next year.
The new wording in the social contract says "We promise the Debian system and all its components will be free." This indicates an ideal state which hasn't been reached but we are working towards. We have not lost any honor or credibility etc. by saying "Yes this stuff is important but right now we're doing a release. We promise we'll get to it ASAP post-sarge." All the GR did was make the language explicit. There is no new sentiment expressed in the social contract which wasn't there before. So if we could weasel out of postponing (not ignoring) the GFDL issue for example then, there is no reason why we couldn't do that now. In fact I bet you we will find a way. [update: SteveLangasek is doing just that.] So apart from general confusion and consternation amongst people who are unfamiliar with Debian and some unnecessary wear and tear on my 'd' key, what did all this brouhaha actually acheive?
-- Disgusted, Tonbridge Wells.
Today (well, yesterday by now) I attended a meeting on Debians' behalf of several New York-area free software groups. The meetings agenda was to try and begin to organize a replacement for Linuxworld which as of next year is moving to Boston. Committees were formed to work on various tasks. Yours truly is doing the vital task of name/logo/mission statement brainstorming. More news as it happens.
While packaging the latest webmin version I had an interesting time trying to figure out how to do stuff to files with names like config-*-linux. The trouble is '*' gets interpreted as a wildcard character. Normally in bash on the commandline to get a literal '*' you would just do this: config-\*-linux. But in a makefile, it gets converted to config-\\*-linux. Increasing the number of slashes was suggested but to no avail. The correct answer is $$'config-*-linux' (You have to use two dollar signs so it doesn't get interpreted as a make variable.)
The second beta of the next generation Debian Installer has been released for the i386, powerpc, and ia64 platforms.
Dan Shearer started a thread in debian-devel about this topic. Here is my list.