David Brook's New York Times opinion piece Stressed for Success? (free registration required) is a good angle on what's wrong with education in America. Brooks is writing to high school seniors who are awaiting the moment of joy (or doom) of college acceptance (rejection). It was hard to pick just one part of the editorial to except (and I don't feel right block quoting the whole thing), but here's the quote that stuck out to me the most.

More than anything else, colleges are taking a hard look at your grades. To achieve that marvelous G.P.A., you will have had to demonstrate excellence across a broad range of subjects: math, science, English, languages etc.

This will never be necessary again. Once you reach adulthood, the key to success will not be demonstrating teacher-pleasing competence across fields; it will be finding a few things you love, and then committing yourself passionately to them.

Just before I left for D.C. I got a Griffin PowerMate. So far I've been using it to relive the stress on my right hand by using it to control scrolling in scroll intensive applications like NetNewsWire (combined mode) and various browers -- it's saving some wear and tear on the fingers that I would normally use for the scroll wheel (fond of it as I am).

I've also started playing with QuickSilver, and so far, it seems to be pretty good. I've only had it installed for a day, so I haven't had to fully configure it, but I haven't found anything that I really miss from LaunchBar yet, and I do like the ability to have a clipboard stack.

I'm slowly recovering from the PyCon experience. I had a great time; the experience was very similar to an ApacheCon. I've been to some of the bigger conferences: JavaOne, Software Development, and OSCon, but there's not the same energy (for me) as PyCon or ApacheCon. I love walking through the halls knowing that people are actively working on stuff together, not just gathering to talk about it. People are excited about what they are working on and eager to share it with others. Score one for the community organized, developer centric conferences.

I found something refreshing about the Python community as well. There's no griping or fighting with Sun or arguing about the (ir)relevance of the JCP. All of that stuff is incredibly distracting, not to mention annoying. The Python folks don't have any of those concerns. Their only concerns are making Python better, and doing cool stuff with it. Sean Gallagher reports that he had several conversations on how the PSF might attract the attention of Sun or IBM. While I'm sympathetic to the desire for validation, attention, and money, I think that the Python community is better off without the large technology vendors weighing in. There are lots of small groups of people working on interesting stuff in the Python community. Once the big guys show up, all of that will be wiped out.

I also found an interesting contrast between the PSF and the ASF. The PSF is fortunate to be focused on one thing: Python. The ASF has gone from HTTP to Java, XML, Perl, Tcl, and beyond. While I appreciate the diversity of the ASF (indeed, when I go to ApacheCon I spend more time hanging out with the httpd guys rather than the Java/XML guys), I've bee pondering about the size. Is bigger truly better? So it was very interesting to see the PSF, which is a younger, smaller organization compared to the ASF.

Marc Canter added his commentary to Liz Lawley's post in which she admits caving in and using an aggregator. I'm glad that Marc recognizes that a number of the properties of a digital lifestyle aggregator will be enabled by Chandler. We still have quite a ways to go before Chandler is ready to be the infrastructure to build something like what Liz envisioned in her post, but reading posts like these keeps me energized about what we're working on.

It wasn't the worst travel situation I've ever been in, but 7 days is the longest I've been away from the family since the girls were born. After 7 days of eating at restaurants or out of a box, I was ready to go home. The only problem was that Northwest didn't want to oblige me. Julie called me during the final moments of Bob's talk to tell me that Northwest had cancelled my flight and booked me on a flight the next day. That caused some whirling until I called Northwest to find out the deal. It turns out that the plane from D.C. to Detroit broke down (Bob Kuehne, did you get home okay?) and when that happens they just book you a flight in case they can't contact you. The nice lady at NWA booked me a new flight on Continental. So I got my bags from the hotel and went over to Continental, where I proceeded to stand for 30 minutes at the ticket counter while the Continental agent was on hold with Northwest. Apparently, Northwest didn't transfer the ticket to Continental, hence the problem. And to make that segment complete, I was selected for additional security screening. The return trip consisted of a 3 hour flight to Houston, a one hour layover (complete with nasty airport food), and a 4.25 hour flight to Seattle. The remaining travel hassle was that it took a while for my checked bag to come up, which made the trip from the airport to the ferry terminal a little tighter than I would have liked. I walked up to the ticket booth for the 12:15AM ferry at 12:12AM (ferry time). Fortunately, the boat from Bainbridge was a few moments late (which the limo driver and I saw from the viaduct), which afforded some breathing room. Otherwise it would have been tight. But all in all, being home now is much better than being home sometime this afternoon.

Unfortunately, I spent more time travelling today than I did at PyCon... more on that in the next post. The winning talks for today were Bob Ippolito's 60 Minutes of MacPython, David Ascher's Flour and Water Make Bread, and Anna Ravenscroft's talk on distutils. I've basically decided that any Mac applications that I write (if any) will be written in Python using PyObjC. There's just no way that I'm going to learn Objective-C -- too much platform lock in for that -- same with Cocoa. David's talk was useful because he's been in a product company setting for a while, and his talk did a good job of listing the things that business people and open source developer need to do in order to work with and help each other. I got a bit confused during Anna's talk, because her slides were going by too fast for note taking (but not too fast to read), and so note taking got me out of sync. But the content did look worthwhile. I also took a picture of Anna speaking (I hadn' t taken any conference pictures even though I've been hauling around the family camera) so that my girls would see the women are computer people too. Ed Loper's talk on the Natural Language Toolkit gets honorable mention for reminding me of chart parsing and Eugene Charniak's CS241 at Brown. I'm writing this post on the ferry back home. There'll be more PyCon posts in the days to come.

It's a measure of PyCon quality that I am so busy that I barely have time to write a blog entry. I've met so many people and had so many interesting discussions. I suspect you'll be seeing a stream of related posts as I recover.

In any case, the highlights of Day 2 were: Guido's keynote, the Pyrex talks, and the doctest talk, as well as conversations with Mitch Kapor, Anthony Baxter, and Jeremy Hylton. A few words on each. Guido covered some of the upcoming items in 2.4 -- I got more out of the way that he interacted with the Python community (even though it was a keynote), and his taste in language design than I did out of the actual features (which you can read about in the notes - a reminder that the crazy Mac SubEthaEdit people are churning out notes on many of the talks. It's passed out of my hands -- I'm only hosting some of the documents on the Powerbook, others have taken over some hosting. About all I'm doing is providing some server space. See the contributor's lists to see who you should really be thanking).

I finally understand what Pyrex is and how it helps with Python performance and wrapping C code. I think that its an approach that can be fruitful for a sizable number of problems. Ditto for doctest. Actually, I'm sort of kicking myself over this one, since doctest was in the standard library and would have been preferable for most of the unit tests that I wrote when I started at OSAF. Fixing this would be a nice way for someone to get involved.

Each of my discussions with Mitch, Anthony, and Jeremy could fill one or more blog posts, so I have to put them off for now. As a final "flavor of the Con" kind of thing: Between lunch and the afternoon talks, Steve Holden held a tongue twister contest -- contestants had to read Steve's tongue twister in 10 seconds or less. Someone at Bob Ippolito's table (it may have been Bob for all I know) typed the tongue twister into a Mac and used the "speak text" service on it.

I hope that you are enjoying reading the conference notes as much as we are producing them. This was my first experience with using SubEthaEdit to take conference notes. It worked amazingly well for Mitch's keynote, but less well for some of the technical talks. The more technical and code oriented the talk, the harder it is to take good notes, sharing or not. Still, when it works, its a wonder to behold. There were about 10 people editing the keynote notes. Some people were inserting content, others were cleaning up text and filling in details that had been missed. Others were Googling stuff that Mitch was mentioning and adding the URLs. Still others seemed to take responsibility for grabbing good quotes. if you contrast the keynote notes with the Nevow notes (which for some reason I only took myself), you'll see the difference between solo development and community development. There's now a core of about 4 or 5 people that are doing the bulk of the note taking work, and the notes being posted go beyond the talks that I'm going to to cover many of the talks from the other track (there are "only" two tracks). I can't and won't take the credit for the notes -- a lot of other people are working hard to bring you the notes. I just instigated the use of SubEthaEdit and am providing a place to host the resulting documents, but the real work is being done by a bunch of folks, not me. For the curious, the workflow is to open a text document in SubEthaEdit, let people edit the document, close the document and scp the file up to the web server. No editing is being done after the talks. It's been a fun experience to (once again) work collaboratively with people that I've never met in person. At the end of the day I suggested that the notetakers meet in real life, and we ended up going out to dinner, meeting each other and getting excited about SubEthaEdit. More on that in another post.

The talks that I enjoyed the most today were Anthony Baxter's talk on shtoom (VOIP) and the much anticipated talk on Starkiller. Anthony's talk hit all the important points and convinced me of the points that he was trying to illustrate related to VOIP and Python. I want to try out shtoom when I get home, and I want to think a bit more about how VOIP relates to Chandler. I met Michael Salib briefly before the Starkiller talk, and got a slight preview of what he covered. He's branched out from just writing a type inferencer (which was already cool) to actually writing a full native code compiler for Python. His work extends Ole Agesen's work for Self, which gives me more confidence that his approach will work. Unfortunately, the code isn't available, and won't be until May, because "I need to graduate". I hope that Michael is able to find a job that would help him to keep working on Starkiller. Of course, as soon as the source is available, people will probably help as well. The PyPy folks showed a demonstration of pypy, which helped to convince the skeptical. I am really happy to see the amount of effort going into Python implementation and I hope that the various projects will learn from and if possible help each other.

The talk that I wished I had gone to was Glyph Lefkowitz's talk on ATOP, which was a lot more related to Chandler than I had thought from the title. Actually, it is a bit of a problem that the titles are hard to connect to the actual subject matter of the talks. I had this problem with Andrew Koenig's talk. In his case, fame (Andrew) lost out to subject matter (I had dinner with Donovan Preston). The content has generally been very good, and via the magic of SubEthaEdit, I can see what's going on in the other track just by looking at their notes as they are being written.

In Chandler related news, Mitch's keynote seemed to be well received. I learned a bit more about his motivations for doing OSAF and Chandler, which was helpful to me personally. There certainly were a bunch of people who wanted to talk to him afterwards, and I saw him engaged in hallway conversations throughout the day. We also had a large turnout for this evening's Chandler BOF. Several estimates put the attendance at around 40, so that makes more than 10% of PyCon that showed up. It was encouraging to see that there's still so much interest in Chandler. The session ran from 9pm to 11pm and would have run longer but we had to bow out. That's enthusiasm. People had a lot of questions and feedback for us.

The most important talks for me tomorrow (today) are the Pyrex talks, because every time I say performance to a knowledgeable Pythonista, the answer I get is "Pyrex".

As a sop for the real time information junkies, I'll try to post the SubEthaEdit notes of PyCon talks that I go to.

Update: bad URL fixed. Sorry for the mess.

Today is the last day and we're finishing up on the sprints. Brian and I mostly worked on writing down his experiences with the sprint. The backup team definitely has backup working, but didn't quite get restore working. Still, this is good progress for people who didn't know anything about Chandler 4 days ago. It looks like this work will be continuing after the sprint, so I'm happy with that outcome.

Ben Laurie show up just before lunch, and a bunch of folks went for lunch and discussed capabilities in Python. During the course of the lunch, we made some progress on the technical/coding aspects of implementing capabilities. I didn't had much to contribute since I've never looked at the CPython internals. I'm unsure about the process for making the case for inclusion of capabilities in some future version of Python. There do seem to be a few distinct groups of people that are interested, so I hope that we can find some way to see this happen.

After lunch I spent some time talking with sprinters Phil Gaudette and Roger Eaton about their interest in Chandler and plans for using it, as well as getting their perspective on how the sprint went and what they plan to do (if anything) next. Phil is interested in continuing on with the dump/restore code, which is another happy outcome. Phil also informed Ben and I about RoundUp -- the nosy mailing list stuff looks very interesting. I can see a lot of use for that.

Jeffrey Harris and I went to dinner together. Jeffrey is the other offsite OSAF'er, so it was good to get to meet him in person (at last) and get to know him a little bit. We had a really enjoyable discussion, including the intentional community that he's a part of.

The talks start tomorrow, and I am going to take notes, but not blog in realtime (I think, I may change my mind). I'm probably going to use SubEthaEdit documents so maybe I'll get some help.

Today Brian Dorsey introduced some of us to a cool PyGame, Pathological. Its a pretty cool looking game written in Python.

Brian and I spent some time pairing on the WebDAV stuff today. Actually, Brian did all the work. I helped dig into how the remote repository stuff works (I was learning it at the same time) but got pulled off into various side conversations and IRC's. Brian was able to get a Python DAV library to work for get/put -- said library needs a bit more work to be truly usable. Then we started trying to fit the DAV stuff into the remote repository stuff. The remote repository works via an RPC like model. When we finally realized this we decided to take a step back, since implementing it using WebDAV would be almost like doing SOAP. WebDAV doesn't have a way to simulate the various repository calls, so you'd be reduced to using as a transport for procedure calls, which effectively reduces to being SOAP. So while it could be done (just to prove that it could be done), I think we're going try to look at better ways of using WebDAV for accessing the repository.

Jeffrey Harris arrived today to join Roger and Phil on the dump/restore stuff. This trio was able to get basic dump working, so they'll be looking at restore next, but restore is more challenging that dump, so we'll see where it goes from here.

Brian's been giving me quite a bit of honest but painful feedback, which is really good. I'd much rather get this kind of feedback early and be able to address it as early as possible. This was one of the goals for the sprint, so we're succeeding there even if it's slower going on the code. A short feedback loop for both positive and negative feedback is an important feature of open source development, and if you want to be successful, you need to be open to all kinds of feedback. Tomorrow is the last day of the sprints, so we'll be trying to wrap up the code projects and save some time to get more feedback.

On the dining/meeting front, We went to lunch as a sprint, and one of the topics that we talked about was the potential of restructuring the scientific journal publishing process using the Internet. I also spent some time talking with Jacob Hellen and Armin Rigo about the current state of PyPy. They appear to be making good progress towards a full Python interepreter and have started a parallel optimization effort. Jacob also described A.B. Strakt's Python product which uses a notification based query mechanism (similar to what we want to do for Chandler) to update the user interface. Brian, Jeffrey and I went to dinner with Jacob, Armin, and Laura Creighton (also of Strakt and PyPy). We talked about how we got started with Python, Emacs vs VI (in a very friendly manner), Python IDE's, life in Sweden, and a whole bunch of other topics. And Laura made it very clear that we were invited to EuroPython in June. It sounds like they are doing a really good job of organizing it and are trying to make it affordable for people to come from the States. She also seemed to think that there would be a lot of European interest in Chandler.

Last night I went out for dinner with Bob Ippolito, Donovan Preston and his SO, and Zach Burke. We ended up at a Singaporean restaurant. After hearing some of Enoch's tales of Singaporean dining, I was curious to try something new. While the noodle/seafood dish that I had was good, I hardly had time to pay attention to the food. Instead, I was paying attention to the conversation that Bob, Donovan, and Zach were having.

By the time we got to the restaurant, my head was already buzzing, because Bob gave me a 5 line (well, more than that) explanation of PyProtocols and why they were as useful thing. Contrary to the name, PyProtocols doesn't enforce protocols, as someone coming from Java might expect. Instead it provides a mechanism for mixing in API to a Python class, and then allowing to you provide the implementation from a separate class or even on an instance by instance basis. By itself this seemed to be a pretty useful notion. I'd seen the name PyProtocols, but haven't had time to look into it. I guess I'd better now.

Donovan works on Quotient, a web based PIM written in Python. Their team is fully geographically distributed and uses their own VOIP software to do twice a day VOIP conference calls. We spent a little bit of time comparing notes on working distributed. The solution of asterisk might work well for having VOIP conference calls with the rest of OSAF (Ducky and I were discussing this in IRC sometime last week), particularly on days when folks are working at home. Donovan also pointed me at QuickSilver, which is a free (an possibly open source) competitor to Launchbar. More on this in the future.

The discussion ranged all over Python, and was really useful to me. Bob filled us in a bit on the Stackless sprint. Apparently the Stackless web page is out of date and there's a version of Stackless that runs on OS X. I was cautioned to wait few months till the dust raised at the sprint settles. One interesting(?) thing I learned is that you'll be able to pickle tasklets, which could be used to simulate continuations (this seems like it would be slow), and could also be used to migrate tasklets to a different machine where they could be unpickled and resumed.

Bob and Donovan told me about this hack that they did at last PyCon which eventually became a part of Nevow. From looking at Donovan's slides, the feature is now called stan, and looks very similar to the MarkupBuilder in Groovy. It's interesting that it's accomplished in Python without the use of special closure calling syntax.

I asked the gentleman to recommend projects that I could look at in order to find interesting/advanced uses of Python. The list included (I may have forgotten some) Zope3, Twisted, PyProtocols, and PEAK. That should keep me busy for a while after PyCon. I wanted to meet Bob in person after corresponding with him a bit electronically. I'm very glad that I invited myself to dinner with the entire group.

Today's quote of the day: "It's great when you open up a repository and there's data in it". The Chandler sprint is picking up a little steam today. The folks working on backup restore have learned enough API to do a walk of the items in the repository, and take a first crack at an XML format for the backed up data. Brian Dorsey decided that he was interested enough in WebDAV to try and figure out how to build a remote repository that uses WebDAV as a transport (we already have Jabber and SOAP based transports). People seem to be able to get work done, but there's a lot of terminology and concepts that people are struggling with.

I've also started to meet a few folks in the Python community. This morning I got into a discussion with few of the core python hackers (our sprint is in the same room as theirs) on future of security in python. They confirmed that the rexec and bastion modules are considered broken. It seems that some people are interested in a capability based approach, while others are a bit more wary and would like a more traditional ACL based approach. One of the big objections raised against capabilities was that they would require a much more intensive rewrite of the libraries to accommodate a capability based style when compared to ACLs. It would really be great to have a security roadmap/plan for Python, but it seems that someone(s) need to stand up and make a firm proposal.

There was also a discussion of the restricted Python implemented by Zope. Zach Burke stopped by afterwards and pointed me at the actual code.

In the afternoon, the core Python hackers discussed the upcoming Python 2.4 release. There are still some decisions that need to be made before there's a real release plan. It was interesting to watch how a release is handled. Coming from Apache, it's a little foreign (read different, not bad) to see how decisions depend on Guido's opinion. This wasn' t a surprise, it was just interesting to see the process in action.

Other discussion that I had today were related to Perforce versus Subversion, and an interesting discussion with Michael McLay that covered ground ranging from funding open source to how to promote Python adoption.

It's been a long day so far. I woke up at 7AM and wandered over the the Cavritz center where the sprints are being held. I was the second person to show up, so I helped wire up the two rooms for the sprints. People started drifting in a bit after that, and in no time at all we were up and running.

The Chandler sprint is going fairly well. In addition to the folks who signed up, Brian Dorsey, who I met at the Seattle Python User's Group, has joined the sprint, and is asking all kinds of interesting questions, and is giving some good feedback on documentation and API designs. Bear from #chandler is here and it's good to finally have a face, voice, and facial expressions to put to the name. He also came "loaded for bear" -- he brought enough backup equipment to run another sprint all by himself. Roger Eaton is here climbing the Python and Chandler learning cures, and Michael Bernstein also joined us for the day. We had some glitches getting started -- mostly because it's been a little while since I've really used a Windows build, and partially because I was barely awake. But we've got things straightened out now, and we'll be ready to start on the actual sprint work in the morning. The most commonly asked question is "Why aren't you using ZODB?". There is a page on the Chandler wiki that contains the answer. Jeremy Hylton did stop by and give us (mostly me) a mini tutorial on ZODB so I could understand how it compares.

It's been good to start connecting names and faces. Itamar Shtull-Trauring stopped by to in-person introduce himself, and I know that there are a few other folks floating around who I haven' t met in person yet. Please stop by the table -- don't be shy.

For all the latest news on PyCon, head over to the PyCon site and subscribe to the PyCon feedpaper.

Taking a flight is always an adventure from the island. As my father likes to say, "Today I travelled by air, land, and sea" - meaning a plane flight, several car/bus rides, and of course, a trip on the ferry. One of the bits that I never like is the trip from the ferry terminal to and from the airport. Option are limited. You can take the bus, assuming that the bus is running, but that frequently results in lengthening an already long trip by an hour or more. You can take a cab - there's always a fleet waiting right outside the terminal. In theory this is supposed to be a good option -- the city supposedly sets a standard fare, and there's no advance reservation required. My experience is that the cabbies never honor the standard fare, and a ride from the airport hovers around the $30 mark. The last option is to call one of the many limousine/town car services. I've used this option with some success -- the key seems to be finding a good driver/firm, which I have. If by chance you are a Bainbridge Islander, try Moges at City Express Towncar and Limo (206) 510-1900. I've been using him for a few years, and he's been very reliable. Today I actually rode with one of Moges' friends, because Moges is still visiting his home country in Africa.

The ride to the airport was uneventful. The checkin process could use some help. I checked in via the Northwest Airlines web site, selected my seat, and printed out my boarding pass. When I showed up at the airport, I still had to stand in line with the other e-ticket people, and then the machine wouldn't take my confirmation number, necessitating human intervention. Good thing NWA is being efficient. I made my way down to the South satellite at SeaTac, and proceeded to start cruising for a power outlet. I've gotten pretty good at finding them in airports. Unfortunately, finding them is only part of the problem. I found an outlet, but someone had plugged his laptop and his iPod into both of the outlets. Not a very polite solution. I think that I need to start carrying a short 3 prong extension cord, so that I can share with others who are doing this sort of thing. Fortunately, I found another outlet. But the seating was on the hard tile floor. Still, AC power is not to be argued with, especially, when contemplating 5.5 hours of flight time and two batteries that last 2.5-3hrs each.

I don't fly NWA very often, and so far, I've been reminded of why. I was summoned to the ticket counter to have my seat moved because they got a different plane. This resulted in me being in the very back of the airplane (but still in the aisle). Upon leaving the gate, the pilot announced that we had to return to the gate due to having 1 person too many on board. He piped up moments later to announce that actually things were fine. The only pleasant thing so far is that I was actually served a meal that I could eat. Getting a meal on a flight noways is rare anyways, but many of the meals contain dairy, which causes problems (lactose intolerance). In the past I've tried to arrange for non-lactose special meals, but that usually results in a fruit platter or a stack of rice cakes. In recent years, I've given up and mostly gambled, but I can' t remember the last time I had a meal that I couldn't eat (this may have something to do with the number of meals provided).

From the second leg of the flight:
Security precautions are still in place: my flight is into Washington D.C.'s Reagan National Airport. The captain told us that all passengers must remain seated for the last 30 minutes of the (1 hour) flight. Failure to obey this rule could lead to us going "somewhere else"...

Finale
I arrived at my hotel too late to order room service or have a reasonable hope of finding an open restaurant. A pair of Clif bars will have to do for a meal. Get assigned a room. Look for an internet connection (always do this before unpacking in case you need to move). No RJ-45 looking plugs in sight. Call down to the front desk. You can use the business center for 20 mins. Ok. That's not working and there's an irate guest in there too. Back to the front desk staff. Oh. Your room doesn't have an internet connection. I ask to be moved to a room with one. The staff obliges, both cheerfully and helpfully. I go to the new room. Check the internet. It's there but not working (the LAN light won't light). I'm almost ready to give up. I decide to call tech support. They say the most common problems are the little black box, or the cable (in between telling me that they don't really support Macs -- I suppose I better get used to hearing this). So back the front desk to borrow one of their cables. Plug it in, and the LAN light finally comes on. Of course, this means my spiffy retractable spool ethernet cable is busted. Last time I used it was when I went down to OSAF in December. I wonder how long it's been broken? I sure hope the wireless net at PyCon is working... Only 7 hours till the sprints begin, but I'm not ready for sleep. Flying eastward always messes me up for a few days, anyway.

In his most recent SuitWatch, Doc Searls gives a great explanation of why I ended up with a Powerbook and not a T41p.

The trusty 1GHz P3 that is the main server for the house has been up for 420 days. Until today. One of the hazards of living on an island is that occasionally the wind comes roaring across and knocks down some power lines. It happened twice today. The UPS kept the machine running through the first outage, but the second one was too close, and the UPS wasn't able to recharge. At least the UPS can talk to Debian via nut, so there was an orderly shutdown instead of a crash.

With the hubris of long uptime now removed, I decided to try and upgrade the kernel from 2.4.19 to 2.6.4. My development server is running 2.6.1 with no problems so I figured I had tested. The next problem was that I tried to get fancy. Not only did I try to upgrade to 2.6.4, I tried to patch 2.6.4 for User Mode Linux, which I've been dying to try. That didn't work at all. The patched versions of the kernel just wouldn't build. So I scrapped that idea and just built a stock 2.6.4. Using make oldconfig which saved a huge amount of time. I wish I had learned about that one a few years ago.

2.6.4 booted like a champ and everything was fine until the system started the firewall. It seems that eth0 and eth1 decided to switch which NIC's they were assigned to, breaking the firewall rules and a host of other interface dependent stuff. After a little googling and fussing, I decided to give up. I'm supposed to fly to PyCon tomorrow, and a new kernel on the machine is just not a smart idea. So I'm going to just leave it the way it is an try to fix it when I get back. If anyone can explain to me *why* the network interfaces switched or more importantly, how I can switch them back, I'd really appreciate a comment.

Time to pack for PyCon now...

It looks like Scoble was successful in getting David Allen to blog. Robert, I'm ready to make good on the beer!

After several months of hard work by all the developers and members of the community, we've released the next version of Pyblosxom.

0.9 has the following changes in it:

• Added a logger and locking code
• Overhauled the xmlrpc system
• Added the verify_installation code
• Added handle callback
• Added Atom 0.3 flavour
• Overhaul of code base to fix concurrancy issues
• Added support for Metaweblog API
• Bug fixes, security fixes, and optimizations

It's highly encouraged that all people upgrade since plugins written for 0.9 will not work in previous versions of Pyblosxom.

If you have any problems, let us know on the pyblosxom-users list.

PyBlosxom main site

Tonight I did a presentation (slides here) at SeaJUG on Groovy. Things got off to a rough start, due to glitches with the Powerbook. This was the first presentation that I've done with it, and I was unfamiliar with the vagaries of Mac PowerPoint and managing a projector. It all started when I tried to hook up to the projector and was told that the example code was too small in the back of the room. I used 16 point unbolded Courier New, which I've used for presentations before (albeit in bigger rooms and screens). So there was a period of futzing with the display management to get Powerpoint on the projector and then getting the projector resolution to 800x600. Next followed the usual SeaJUG introduction and rituals. After I was introduced I discovered that I couldn't unlock the screensaver -- no matter how much of my password I typed, the dialog only showed three characters. Perfect timing. I just pressed and held the power button. After the machine rebooted, I was able to log in and get going. (Perhaps I need to get a copy of Keynote...)

There were lots of questions, some technical and some philosophical. I tried to just explain the features in Groovy to people as well as potential benefits. There was a group of people that really liked what they saw, and there was a group of people who really didn't like what they saw. I expected this, because to fully buy into Groovy, you have to buy into some assumptions about how software should be developed, which is philosophical bordering on religious. Not to mention that it's a whole talk by itself to do it right.

In addition to the philosophical questions, there were a lot of reasonable and sharp technical questions about various aspects of the language. I wish I had had a tape recorder so that I could have gotten all of them, as feed back for the Groovy development team. On the whole, I felt pretty good about it. I was a little less prepared than I wanted to be, but people didn't seem to mind.

One bonus for the night was getting to meet Kellan Elliott-McCrea. Kellan needed a ride to the meeting so he (and I) hitched a ride with Wilhelm.