Category: Uncategorized

After several days full of complicated IRC discussion about who is driving whom where this weekend in the tangle of people going back and forwards between Yass, Canberra and Sydney I bring you the wedding car pool. Get it while it’s hot.

Ubuntu 5.04

Released today. I upgraded my servers this afternoon. (It’s a personal thing, I love watching aptitude run.)

I’ll be interested to see how long I resist the temptation of upgrading to the next development branch this time. (For Hoary, I waited until upstream version freeze I think.) For my uses, Hoary hasn’t seemed to be a significant improvement over six-month-old Warty, although I’m told under the hood that it’s much nicer. Perhaps on the next cycle I’ll wait until the actual release… unless suspend starts working, of course.

Smarter than your average stick

During the course of the Twisted Sprint, Mike made a very useful suggestion about our documentation: that it should begin by defining its target audience. I’m going to try and do this across the existing documents eventually. Failing to do this in the past have resulted in some spectacular misunderstandings of when you can use bits of Twisted (at least one person has appeared on IRC thinking that enterprise is usable without knowing how to write a SELECT statement).

Later on in the weekend I briefed Jonathan (read: ranted at him) about what we need to do about the Deferred documentation with regard to targeting it at users who have never encountered asynchrony before.

The original documentation seems to have been written revelation style: gather unto me my friends, and I will free you from the heavy chains we have all laboured under trying to manage our own callback sequences: I give you the Deferred. (Cue weary lifelong developers crying in relief.) The actual users of the documentation (at least, the ones who complain about the docs) seem to tend more towards the “callbacks? wtf?” audience than the “at last the scales have fallen from my eyes, lead me forth” audience.

The upshot of this is that we probably need extensive documentation on the fundamental concepts of asynchrony. So far so good.

And then, just this week, a post to Twisted Web threw me another curve ball (I don’t know what the equivalent term would be if I were making a cricket analogy, sorry Andrew): Twisted (or Twisted-ish things like Nevow at least) have people pottering around with them who don’t really know much about concurrency. Or, quite possibly, don’t know anything at all about it.

This can be reduced to a target audience problem eventually: at what point does the documentation take the smarter than this stick position?

This angst brought to you by one semester’s worth of memories from teaching second year concurrent programming and having people ask each other in the tenth week of the course, completely mystified, why their GUI was freezing. Do get in touch if you want to write a pre-pre-Deferred document (or was that post-post…?) that is willing to discuss threading in detail, since, yeah, people will have to know when to defer to thread at some point or other.

If anyone would like a lift to Jeff’s and Pia’s wedding next weekend that fits into the following schedule, let Andrew or myself know (if we get full-up, I’ll update the entry appropriately):

• Saturday 16th: Leave Sydney mid-morning and travel to Canberra. I think we’ll stay in the City Walk Hotel if they have room, but could drop you off elsewhere.
• Sunday 17th: Leave Canberra mid-morning to travel to Yass in time for the ceremony (going straight to the venue).
• Sunday 17th: Leave Yass in the early evening to travel back to Sydney (we’re not going to lca and thus will not be returning to Canberra). This will either be straight after the reception, or very soon after the start of the party.

If you only want a lift one way, that’s fine too.

Note especially that Andrew is driving the entire way. He’s currently on his learner’s permit, and needs to accrue more hours in his logbook before his driving test. This has a couple of effects on our passengers aside from his driving (he’s pretty good, btw, and the Hume Highway will be no problem for him after the Bells Line of Road over the Blue Mountains). The trips will be slow because learners are speed restricted to 80km/h. You get not one, but two designated drivers, since I must be under 0.05 to supervise him. (Andrew won’t be able to drink during the entire weekend, because he, as a learner, must have a 0.00 reading.) But if you’re happy with a completely sober advanced learner driving you about, get in touch.

I, like a lot of people, don’t have the art of taking photos of people hacking and making them interesting to look at. But, hey, photos!

It’s been a good weekend so far. Chris, Andrew, Andy and Tim really have been working on a virtual firesystem abstraction, although without the April 1st insanity. Last I heard the aim was to have a demonstration server which allows (virtual!) files to be uploaded and viewed over SFTP and FTP and viewed through Twisted Web at the same time. Andrew has been buried in the guts of FTP: the man has a long burn out time but he’s back in the game. I asked him for some attention about two hours ago and I’m still in the queue.

I added a few tests to bugs yesterday in an attempt to learn some trial, and converted a few demonstration code bits to automated tests. Today I’ve been much more idle than everyone else, but I did make a start on a documentation writing style guide that could keep me busy for years if I make the ropes tight enough.

Twisted sprint

Andrew and I are flying to Hobart this evening for the inaugural Australian Twisted sprint. Sprinting will be competing with tourism a bit, but I’m looking forward to both, despite this being another weekend in which I am not sponging around on my beanbag.

Google spam, round 3

Seems Google’s business development people (or their cyborg stand ins — the mail has a reference number, so it’s possible it’s all handled by The Machine) were concerned I didn’t reply to their last mail. So, they sent a followup.

I sent the following in reply:

 On Wed, Mar 30, 2005, Google Crawl Coverage wrote: > Greetings, > > I just wanted to follow up with you to confirm that you had received my > previous email. If you are not the right person for this, perhaps you > could forward it to the appropriate person within your firm. > > Thanks. I look forward to hearing from you.  I received it. I've deliberately chosen to have a number of domains I run, including this, excluded from Google's index (and from many other robot crawls). It's not a site designed to be useful to casual viewers, and it isn't a commercial site so attracting extra visitors has no commercial advantage to me.  I'm concerned about this particular business development scheme of Google's -- what if every search engine sent me repeated queries about my robots.txt file? That would be very time consuming to deal with. I know the implications of its contents and deliberately chose them.  -Mary 

And hopefully that’s the first and last time I’ll reply to spam. Invoking the categorical imperative in that reply was silly anyway: much of the idea of a business plan is to do something that no one else is doing and hope that it doesn’t become a universal law.

Twisted 2.0

It was released. And that‘s exactly what I would have said about it too. So handy to have people to take care of these things.

More on conferences

The revival of the "why do you/don’t you go to tech conferences?" discussion on LinuxChix reminded me of my response last time: an article on making your conference website accessible to the conference-naïve (who aren’t necessarily undesirable speakers). If you’re interested in attracting proposals from any group that isn’t currently represented (be they women or QA people or whatever) your conference does need to make a special effort to aim its website and all other publicity at outsiders. Insiders just need the conference dates and the closing date for the Call For Papers. Outsiders need everything spelt out: what’s a ‘paper’ in this context? how long is a talk? what’s the selection process?

Since there’s nothing better than a topic which allows me to be both on-topic and narcissistic, I thought I’d explain my perspective on conferences in the light of the "women at tech conferences" weblog-thread:

• Liz Lawley: why sxsw?
• Shelley Powers: Number 9 Number 9 Number 9
• Kathy Sierra: My first ETech comments (via Dorothea Salo on misbehaving.net)
• (and, finally, a conference-I’d-actually-like-to-go-to post) Andrew Kuchling: Musings on gender balance

After all that lead up, I’m sorry to have to confess to say that my complete failure to attend a technical conference over the last year has nothing to do with their gender balance at all. The most I can contribute is to say that all other things being equal (which they never are), I’d probably have a better time at a conference approaching a 50-50 balance than I would at one approaching either extreme. As it is though, if I want mixed gender or female dominated spaces I know where to find ’em.

There’s two tech conferences I would have been tempted to attend this last year: linux.conf.au and PyCon. (linux.conf.au is, um, 3% female in attendance and around 5-10% in speakers I think. PyCon I don’t know about.)

The biggest single reason I’m not at these two conferences is holiday time. If I want to attend either of them, I need to take annual leave. Back in the days where I was unemployed and/or a student, a week at a conference did not automatically mean spending one week less by the sea that year: now it does. (My employer, like many, does not have its employees attend only peripherally relevant conferences on company time. Fair enough.)

This makes me a big cowardly custard in Free Software terms. (I know a bunch of people who would never consider having a holiday that didn’t involve a tech conference.) But there we have it. If I have to trade my beloved time-by-the-sea time off against a conference, it has to be a bloody good conference.

There are a few specific points that stop each conference qualifying as a ‘bloody good conference’. In PyCon’s case, it’s almost entirely the location. It’s an expensive trip from here: in addition to ‘spending’ holiday time on it, I need to spend thousands of dollars on the airfare. In linux.conf.au’s case it’s two things: I don’t drink heavily enough to find the after hours stuff all that appealing and I didn’t enjoy many of last year’s talks. (I don’t want to lean on that point heavily: I think it’s a sign of a mismatch between my interests and those of the program committee, rather than the committee’s choices being bad.)

What of women at conferences then? Well, some of the absent women are just stingy with their holiday time and their finances, just like, I presume, the absent men. There probably are some factors that specifically stop women attending and presenting, but I don’t think they’re anything compared to two things: women who are professional techies are less likely to have it as a hobby too (shades of my time-by-the-beach rationale here I admit); and there’s a huge starting gender imbalance in techies. I’m therefore not finding this particular issue very stirring.

<spiv> Ok, another upgrade, this time for universe. This should be the last net connection-choking thing I do today.
<spiv> Uh, weird. It self-aborted mid-download.
* spiv lets aptitude just do whatever.
<spiv> Ah, fuck.
<hypatia> ???
<hypatia> *suspense*
<hypatia> ???
<spiv> dpkg segfaulted, and now the machine has hung.
<spiv> *sigh*
* spiv sysrq reboots.
<hypatia> Ew, suck
<hypatia> Maybe you should have done the whole thing at once??
<spiv> Possibly, but I wouldn't think so. Anyway, I'm fscking to make sure it's not like the weird error I saw on trogdor.
<spiv> Anyway, file this one away for next time you feel like computers only hate you 🙂
<spiv> They hate everyone 🙂

Consider it noted.

Later:

<spiv> Ah, kernel BUG during fsck.
<spiv> I'll try without the proprietary NVidia module, then I'll try vanilla 686.

Just for reference, he’s trying to upgrade an underclocked (long story) Athelon machine with a August 2004-ish snapshot of Debian sid to Ubuntu Hoary (as of today). Um, don’t try this at home… ?

Is this really Google spamming me because LinuxChix Live‘s robots.txt file is a universal block?

This line at least does seem to imply that it came from their netblock (the public google.com IP addresses also begin with 216.239):

 Received: from 216-239-45-4.google.com (216-239-45-4.google.com [216.239.45.4]) by fuchsia.puzzling.org (Postfix) with ESMTP id 4D2DF1016C for <additions@live.linuxchix.org>; Tue,  1 Mar 2005 05:58:08 +1100 (EST) 

That’s kind of … uncool. One might almost say… unsolicited commercial email.

I did get something like this once before and forwarded it to abuse@google.com, only to get an instantaneous automated reply telling me that any spam I was reporting could not possibly have come from them.

Perhaps it didn’t and it’s a rather more competent header forgery than most spam. The forensics will be interesting in that case.

I bring you this list of things I do to stay sane on the internet more in the spirit of catharsis than anything else. In saying it, it no longer binds my heart. I am free! And so on.

By no means is this intending to be convincing: you do not have to choose your reading material like I do. By all means read things that piss you off. Even more important, read things that piss me off. Someone has to do it. I choose… you.

That said, I bring you Mary’s rules for staying sane on the Internet:

1. re-read Charles’ Rules of Argument monthly;
2. do not syndicate an RSS feed if anything in it ever causes me to stand up and pace around having an imaginary argument with the author;
3. do not syndicate an RSS feed if more than one-twentieth of the posts are commentary about weblogging as a fabulous community building tool (and that number is only one-twentieth so that Crooked Timber and Easily Distracted slip under the radar), making exceptions only for people in their first months of weblogging;
4. unsubscribe from an RSS feed when I read less than half the content;
5. permanently unsubscribe from all mailing lists on which I read less than half the content;
6. temporarily unsubscribe from mailing lists when at least two posts in any one thread cause me to pace around the room having an argument with either their author or everyone who’s contributed to the thread;
7. permanently unsubscribe from a mailing list the third time the previous condition is met (I make exceptions for the LinuxChix lists because they’re the only very active and vibrant social lists I’m on); and
8. delete unsent about one third of all emails I compose to mailing lists.