Older blog entries for Rasputin (starting at number 10)

Another snowy Sunday. I really suspect I should be living in Barbados. At least until global warming improves things in Ottawa.

Learning is never fun ;) I just learned I need to find somebody who knows more about device drivers in FreeBSD than I do and is willing to answer questions with more than the usual "just look at the source from an existing driver and go from there". Anybody in the Ottawa area that would be willing to help, I'll buy whatever your favorite drink is for an opportunity to pick your mind ;) Otherwise this will probably take substantially longer than I originally thought. Such is life ;)

On with the hacking adventures...

Saturdays are obviously a poor choice (at least in my life) for hacking. Working on device drivers with a hangover is probably not the best idea I've ever had.

Had an interesting get together with Baba (my SQL and standards Guru) and some of the people he works with last night. Terry was there as well (although typically late) and we had an interesting chat about why most developers don't seem to want to understand the concept of meta-data. We also chatted about the impact poorly trained money chasers (MCSE. CCNA, etc) are having on the industry. He was amused by the story about the CCNE that set up a DNS as master for the "." domain then couldn't figure out why it wouldn't forward queries.

I finally got the base Travan driver to compile. For my next trick, watch while I pull a rabbit out of my hat ;) Actually wasn't that hard, it was based mostly on the QIC driver from the 3.x stream. Now I have to sort out the underlying connections so it can actually do something. I want to have it happenning tomorrow, but I think Mon night is probably more reasonable.

My wife picked up a copy of Phantom Menace for me (which covered last night until way too late). I felt almost as excited as I did in 1977 when I first saw episode 4 in the theatre. I can't believe it was that long ago. That movie was a real part of my inspiration in getting into computers in the first place. I even spent some time building industrial robots in the early 80's because of that movie (that was actually not the best choice I ever made ;)

Vote the Meadow Ticket !

Another day, another $0.70(CDN) after taxes.

#ifndef SKIP_RANT

Why is it that when a new printer gets plugged into the network everybody feels obligated to print a test page? One is more than enough, everything else is paper airplanes.

#endif

On the subject of printing, there has to be a better way. I will admit to being a big fan of "circles and arrows and a paragraph on the back of each one explaining what each one is to be used as evidence against us.." (with apologies to Arlo Guthrie ;) but it's a major pain cleaning off the monitor...which leaves only the tree-ware.

One of the team here found a new (new to us, apparently it's been around for a while) network tool, called dig. Wanted to know more about it so we checked for a man page in Solaris...no such beastie (although that's what he was running it on). Checked HP-UX and same response. Just because I'm like that, I checked FreeBSD and I found it ;) Mark up one point for the open source docs! ;)

Zero accomplished yesterday. Staring at a monitor with a sinus headache just doesn't do it in my little world. Hopefully I'll get a bunch done tonight after a visit with my 'SQL Guru' at his preferred place for Irish Pints. I'm always amazed at the depth of this man's knowledge about SQL and programming language semantics in general. I keep trying to get him involved in open source (even though he's not a hacker) but he feels that standards are enough of a contribution.

Bill and Opus in 2000

I really feel as though the planet-wide clue-meter shallowed out while I was sleeping. I saw another post in one of the lists I follow (FreeBSD-Hackers, I think) from some would-be journalist asking us to explain why FreeBSD is better than Linux for his article. This question is not only not answerable, it's completely meaningless without a context - FreeBSD is/is not better than Linux in what circumstances, what way... You have to look at specific needs and circumstances and make a decision based on that. Even Win (please don't get your knickers in a knot, I'm still an open source type guy) is better than BSD or Linux in some situations. Specifically, for the average home user with neither the time nor desire (and arguably no ability) to go beyond starting the box, reading email and browsing the web (or playing games) the burden of learning and administration of anything except Win is probably excessive. That will probably change in time ( I suspect the majority of people in Advogato are working to change that ;), but it is the case right now.

I get into work today to find customer questions that I would never have expected to see. How can someone who apparently knows so little about network administration have a position of responsibility over a large corporate network? The only good news about it was that it moves me back to the end of the queue for cases, so having resolved the problem quickly, I can get on with important things (like figuring out how to write a BSD device driver ;)

To make absolutely sure I was paying attention, I actually heard there is some consideration of making bridge a demonstration sport for the 2002 Olympics. Complete with random tests for performance enhancing drugs. I really wish I was clever enough to invent this stuff as a joke ;) How is sitting around playing cards a sport? What kind of drugs could possibly enhance your performance at bridge? Steroids? Valium (or would that just help you sit through it)? Caffeine?

I actually have a big piece of my Travan driver for BSD done. I still have a lot of work and a bunch of implementation type decisions to make before it sees the light of day, but it should be ready for testing outside my boxen in a week or so. If I was young, single and still able to function on 2 hours sleep per night for long stretches I would probably already be done. This is a nice (but accidental, really! ;) segue into...

My thoughts on hackers and relationships. It's impossible to generalize enough to say what type of person is most appropriate for a relationship with a hacker. I've been very happily married for 9 years, and my wife will never get beyond very basic computer use (she's not techno-phobic or anything, it's just not something she wants to do). She has, however, helped me grow spiritually, which I never really would have done on my own. We are complementary to each other and, for us, this works really well. The only draw back is that I've had to reduce my total hacking time to make sure there is room in my life for the relationship, including our children. Add to that the fact that as I've gotten older I need more "down-time" and my productivity as a hacker has definately decreased since I started 15 years ago. I still manage 2 or 3 hours per night during the week, so I'm not a total write-off. I even do an occasional all-nighter during the week-end ;) My point (other than the one between my ears) is that we're individuals and the solutions to these problems have to reflect that. What works for me is not guaranteed to work in all cases ( or any case as far as that goes) and what works for someone else cannot be assumed to work for me.

This reality brought to you by Kafka, illustrations by Dali

Some days it pays to be a pessimist :/ I'm not particularly amused by waking up to see snow in April, but at least I wasn't unpleasantly surprised.

On the subject of the Dimwit cert, I was looking at that as more of an indicator of who gets to wear the pointy hat ;) We all occassionally do Silly Things (TM) and this cert is a good way to highlight them. I suggest, however, that it should be a strictly temporary thing. Maybe expiring after a few days.

My fingers apparently don't want to cooperate right now. I guess the third coffee is needed then off to the book store to see what new bits of nearly useful information are available.

Spent an interesting couple of hours in the Mess (I left the Military after 10 years to return to my Geek habits ;) a couple of nights ago talking about what makes a company enlightened in it's handling of employees. I guess, by most standards I work for a pretty good one. I've certainly worked for some fairly nasty companies in the past. Here, they didn't even complain when I took an older box and flashed up FreeBSD 3.4 on it. It doesn't do a lot yet, beyond basic desktop functions, but a couple of us are trying to convince the powers that be that we should be allowed to port some of the software around here to it. It's already better behaved with our products then a platform we support (guess who...)

I noticed that wrong information on a mailing list (no matter how well intentioned) , although amusing, is probably more dangerous than no information. It's not nice to panic your customers, although this can provide "hours of entertainment for the children" ;)

Judging from the number of diary entries in general and the specific comments about the volume of traffic, I would guess we're about to hit critical mass for something to be done about scalability. I'm not sure how this should be approcahed, though. It's interesting to read the other diary entries and occassionally comment. The best approach will probably have to involve some type of personalized filtering, maybe based on who's diary entries you like to read. I'll definately have to look at the code for this site and give this some thought. I hope (would bet) that a large group of people are doing this already.

Another typical spring day in DisneyLand on the Rideau Canal. I much prefer sunny and warm, but that's still a ways into the future. Since I'm going to be staying indoors all day, I have no excuses for not getting a bunch of stuff done. We'll have to see what happens.

I finally broke down and cleaned out my "junk box" and discovered and old HP T1000 with some tapes. Popped it into my FreeBSD 4.0 box and lo and behold, it didn't work ;/ Further investigation reveals this puppy was never supported under *BSD, although the QIC series was until 3.4 or so. So, I spent the day (when I wasn't chasing the kid around the house ;)) getting started on a Travan driver for FreeBSD. Of course I'll get the driver written only to find out the drive was in the junk box because it belonged there. Such is the life of a hacker.

I've actually given some thought to what I was thinking about yesterday, an article about the international standards process. I suspect there might be some interesting lessons in that particular process for the open source community. I'll pop it up for comments when I get to about the third revision and we'll see where it goes from there.

I was also thinking about the frequent occurence of "hot air" wars (they don't often make it to flames) over the most amazingly trivial things. I really believe not having to discuss these things face to face allows a level of aggressiveness you wouldn't normally see. Most of the hackers I know are relatively soft-spoken and articulate yet put a keyboard in front of them and hit a "Holy Issue" and the pilot light kicks in... I guess that's a feature of the on-line community that will be with us for a while. If there were a way to personalize the process a little more, this probably wouldn't be an issue.

Ain' t it amazing what a little sleep can do ;)

While poking through the daily diary entries, noted a couple of things that need commenting on. First a minor correction...

Schoen - the FreeBSD-QA mailing list started just before we shipped 3.4, although it didn't get real attention until just before 4.0. There had been a lot of whining about the lack of solid testing for stable, one thing led to another and there we were ;)

Asmodai - Having sat on the committe that wrote SQL-99 (ISO Joint Technical Committee 1) I guess I should explain a little about why the storage mechanism for pre-compiled stored procs isn't covered within the standards. Standards have to be written to set a minimum level of support and compliance without imposing (where possible) limits on extensions and improvements. We felt (and still believe) that specifying a language should not include forcing an implementation. By enforcing compiled compatibility across vendor platforms, we would effectively be forcing one vendor's methodologies to be the standard. The reason stored procs are so efficient is they can be optimized to take advantage of the underlying implementation. The optimizations for Oracle are substantially different from those for Sybase because the underlying database engine is substantially different. These products also store pre-compiled procedures differently. This approach also allowed us to permit vendor extensions to the language (PL/SQL or iSQL) while ensuring a minimum level of cross platform compatability (mostly enforced with the fips flag).

There's actually a lot more to the problem than I've explained here, so maybe I'll do an article on the standards process over the next couple of weeks to examine this further. Feel free to pester me about this if I let it slip a little ;)

I'm having second thoughts about a person to person messaging system in an environment like Advogato's. Is there any desire for something like this in the community at large? Any thoughts? My initial thoughts on this are that a wholly self-contained messaging area (within Advogato) would eat tremendous amounts of resources and quickly become unmanageable. An e-mail link (possibly hidden unless otherwise flagged by the owner), much like in SourceForge might be a slightly better idea. It would obviously have to be done outside the context of "mailto:", else we force a specific subset of browsers to use it. I'll give this more thought and continue this <babble> tomorrow.

Give me liberty (from SQL) or give me a beer! ;) It has definately been one of those days. Fortunately it's also the last one of this week.

I finally have my FreeBSD 4.0 box to the point where I'm happy with it, so I guess I'll start poking at some of the PR's to see what I can fix. I still have to decide whether I'll be better able to work on 3.4 (I'm slightly more familiar with the code) or 4.0 (it's new enough, the bugs should all be relatively shallow, because nobody's had time to be really weird yet ;))

I noticed a couple of things that might be worth investigating (I know, only been here a few days and already I'm whining ;)). I realize it's there in front of me and my sydlexia prevents me from seeing it, but there must be a better way to get to the journal entries area so I can babble at my screen. Also, wanted to drop Pedro a quick note about his HP toy and realized I actually had to work at sending him a limited distribution type message (I'm sure not everybody here wanted to read it). Maybe some kind of messaging system? I guess I have to learn XML so I can look at the code that runs this to see how, if at all, this might be done. Amazing how many opportunities to code show up in the open world when you're looking for something to do.

I think I just accidently accepted a pair of tasks...

I was going to add my $.02 (CDN) to one of the articles when I realized I can't post anywhere except here until I get certified a little higher. The joys of being new to a community. This is probably a good way to keep the s/n high, as opposed to what I've been seeing on /. recently. There really needs to be a better way to limit the trolls, idiots et al ;)

It also occured to me as I talked to customers, that there is a lot to be said for making developers actually support their creations. In the corp world I'm used to, I coded and some other poor sucker^H^H^H^H^H^Hspecialist received the concerns and complaints from the people who used it. It was a lovely little bubble I lived in. Now that I'm supporting software written by somebody else, I begin to get a better appreciation of why solid coding methodologies are so important.

I would guess that having to support your creation in large part contributes to the quality of open source software. Knowing you're the person who receives the rantings of an irate user is actually a decent motivation to write good code. There are very few OS projects out there where the support people are a different group from the developers. You also get a better feel for what's good and bad with the product, outside of marketing pressures.

I get to find out how much I've learned as a large customer needs to make some substantial changes to the product we support (they want to change the way their network is partitioned) and I (was) volunteered to do the scripts and code to make the changes for them. If it doesn't work as well as they would like, I get called directly ;) It's not actually open source (my employer isn't that liberal) but the theory is about the same.