Rhonabwy

Developer article at Apple on OCUnit

Apple has posted a nice walkthrough article on test driving your code with OCUnit. Of course, I'd rather it was showing off UnitKit, but shoot - anything to spread the word on unit testing your code! (and OCUnit is a very mature and likable framework).

that terrible clean sheet of paper

Most of today has been a strange iterative battle with myself. At the edge of my brain, I have some thoughts attempting to resolve themselves into an idea - something interesting to code, make, etc. That's not all that uncommon, and I usually just write down whatever interesting idea it is, and when I do I get some relief from the idea. If it sticks around a while, I know its something more than transient, and maybe I do something with it, or maybe I just log it down in my pile of paper journals, waiting for when it seems more doable or appropriate.

Only today I haven't been writing anything down. Instead, I'm sort of endlessly opening up the laptop and staring at it like it's some terrible clean sheet of paper. Endless possibility comes up and sort of mentally drowns me. I stare at the little icons on the left side of the screen (where I keep the dock). They almost seem to be accusatory in the plain breadth of all that I could create with them. Text editors, information tools, programming, web development... so many things.

In between fits of clearly mild insanity, I read up on PHP and looked around in the guts of TRAC and Testlink to see how things were put together. I'm thoroughly engaged and impressed with Trac. It's relative simplicity, and it's incredible reach through several pieces of software, pulling them all together so nicely. That's one seriously elegant bit of thinking in there. Testlink isn't so elegant as much as it really does a good job at scratching an itch that I have, and I was curious how it goes together. I've seen some impressive things built with PHP (phpMyAdmin continues to be my all-around favorite), and it's sort of ironic that I heard someone say the other day: "yeah, it's got a library for freakin' everything, but I'd never build a production system with it". Ironic because I remember attending some sessions at OSCON nearly two years back where I learned just how much production stuff was being done with PHP at Yahoo, and thinking "Damn, I clearly missed the boat on this one!".

Like the endless possibilities on a blank piece of paper, the programming languages and libraries seem to abound with so darn much potential that it's almost overwhelming too. Ruby and this whole Rails things that a number of folks are raving about, python and it's endless potential as it seems to be embedded in freakin everything, good old perl and the ceaseless churn of CPAN with it's libraries for everything and anything, and more recently for me PHP with it's depth and breadth and reaching into the world of web applications and object oriented bits.

I started a project nearly a year ago in python that drove a simple web application. I was cheerily adding and building onto the whole kit up until I started at CoCo, and then it all pretty much slid into a stage of paralysis until I came back up for air. Now looking back on the kit a year later, I'm wondering if keeping the game in python makes sense, or if I shouldn't try and rebuild and retry the whole thing in PHP, or Ruby/Rails, or even go back into Java and develop some servlet thingy. The disgusting thing is the project was very much like Ta-Da List, a small outgrowth/subset of Basecamp, one of those "Check it out!" examples of a success with Ruby on Rails. It did more, and slightly different, things that Ta-Da list, but the similiarity is disturbing. (Gus, I can count on you understanding)

No more comments

I'm just freakin' disgusted with online comment spammers. I've held out and held out, liking the service for where it made me new connections and allowed my friends to post tidbits to my blog.

No more.

After cleaning up 932 comment spams this morning from a certain unnamed not brick and mortar casino, the whole thing just really pissed me off. The templates have been edited, the site rebuilt, and the configuration files all twiddled to just not allow the damn thing.

I'm pissed off that there's such abuse. I'm not really pissed off at anyone in particular, just pissed off that it's come to this.

interview with Wolf on drunkenblog

I noticed via Gus' delicious links that there was an interview of Wolf posted on Drunkenblog. I read it, enjoyed it, and then went to check my email where I found a note from Gus telling me about the interview I'd just read. Sort of ironic that I found it through the "social bookmarking" first.

Wolf is quite the character. Can't say as I really know him, but I've met him at WWDC and I periodically read his blog, which often has neat MacOS X coding tidbits tucked away in it. I remember quite a bit of commentary in the Q&A sessions at the last WWDC on his code, as well as the general fervor and dinner commentary it engendered at the WWDC blogger dinner last year. The attendee list for this coming year is really stacking up, but I haven't a clue how I'd be able to afford going to WWDC again this year.

Pragmatic Studio

I'm lagging other folks in doing that "Hey, lookee!" link thing to Pragmatic Studio, but here it is none the less.

I've never met Andy of the trio, and only briefly met Dave Thomas (who signed a couple of Programming Ruby book for me), but I actually spent more than a few seconds talking with Mike Clark at the last WWDC. And it's really tempting to see if I can swing the $600 fee for their talk that will be held here this summer. I have quite the pile of pragmatic books, but I bet I'd still learn and thing or three from these guys and how they're doing their class.

XCoder meeting this evening

We had an XCoder SIG meeting this evening, and even with a very loose format, the meeting went pretty well. We had decided to dig in to the Aaron Hillegaas book: Cocoa Programming for MacOS X and start working through the chapters to get people up to speed and messing with programming in Objective-C. For tonights meeting, we said "Ok you guys, let's all work through the book ourselves up to Chapter 5 and then convene at the meeting and discuss any of the problems we had".

That all worked out pretty well. I was ready to skip right over chapters 1 and 2, but George restrained me somewhat and we more or less flipped through the book, calling out sections, and asking if someone had a question or problem with any of it. We had some good questions, which I don't think we horrifically misguided anyone on, and gave some previews here and there to topics that would be coming up later in the book. For our next meeting, we're going to work through chapters 6 through 8. And there's been enough interest to have a little weekend "hackathon" centered around doing the book examples and "challenges", so we've set up a time to play with that on April 16th. That should be a good time.

Our attendance was done a touch - we had 13 folks this evening, but I'm feeling pretty good about a steady baker's dozen worth of folks at the meetings. Our peek meetings have run at 20 to 25 folks. I don't know what the CocoaHeads group on the east coast is getting, but it seems pretty darn good to me.

SciTE

You find some interesting things when you're not looking for them.

In the past week, I dived in, through, back out, and circled around several avenues of getting some reasonable load numbers on an internal component at work. In those various permutations and ambulations, I installed Watir, and with it Ruby 1.8 for Windows. For Ruby, I used the One-Click Windows Installer, which sounded like a pretty darn good deal to me.

What I didn't realize is that it had a few extra pieces along for the ride. Mostly, I get really freakin' annoyed with those pieces. Like Netscape 7.2 installing all that crap that you then have to go deinstall. But this time there was a real gem buried in there: SciTE - a text editor for windows.

It's not the "Oh lord this rocks!" text editor that BBEdit is, but when you're stuck in Win32 land, it's pretty decent. I mean, there's always emacs, but I was never really into Lisp and I didn't really want an entire OS just to edit a few files. Now this SciTE is really doing the trick well. As an editor, it's pretty darn nice. And it's got a few features that have made digging into Ruby a lot more enjoyable (like running the current window as a script in a secondary pane).

So even if you're not out dorking with Ruby and learning a little something new, and you're on the Win32 platform and looking for a reasonable text editor, check out SciTE.

Bookworming

I've been spending my evenings doing that Seattle favorite (and mine from long before I ever moved to Seattle): Bookworming.

I finished out Greg Bear's Legacy and Eon. Legacy is relatively new, but Eon's been around for a while. I enjoyed them both, although I think Eon won out of the sets. All told, I'm liking Greg's more recent work a great deal - Darwin's Radio and Darwin's Children are both on the in-the-bedroom-pending-future-rereading shelf.

And now Karen's got me hooked into a vampire novel. I guess every fantasty writing seems to have to do one of these, and Robin McKinley's Sunshine is pretty darn compelling.

So other than doing things like cleaning the kitchen and making dinner, it's been book time. And now that I'm settled back down with a cat on my legs, it's book time again. This time with an apple to munch on while I'm at it.

success, and gasping for breath

I get to claim success for the planned learning/technology catch-up today. And with that, I'm now stunned and mentally gasping for breath. I moved through a whole lot more than I expected.

I've now had a turn at generating HttpUnit tests, wrapping those same kinds of tests up at a slightly higher level with Canoo Web Test, ran into some interesting javascript errors and problems with the 1.7 development release of the Canoo web test framework, reaqainted myself with Grinder (which means a thorough brushing of python since it's inherently Jython under the covers), stepping up even higher on the abstraction plate and using Watir with Ruby to drive IE interactively, and I broke through some "how the (#^%( are you supposed to do that" with IDEA.

So now the summary, for the geeks out there.

IDEA was periodically darned annoying today. Switching back and forth from Firefox to IDEA would end up with IDEA hanging for a few seconds, and then I'd be stuck sort of waiting for it to figure out what the hell it was doing. That being said, the IDE worked pretty nicely once I started getting up to speed with the keyboard shortcuts. Switching between the laptop's keyboard and regular keyboard will be keeping my busy for a while though - at least until muscle memory starts to kick in a little better.

I was incredibly impressed with WATIR and driving IE through Ruby. Using IRB to interactively drive it was cool, even if I did periodically segfault it (or whatever it was that caused it to crash out of the interactive window). From a super-high level, its fantastic, because it's using a browser instead of just emulating it. Making sure all the image componentry gets loaded it one of my key concerns, and this resolves it. Plus, I can get some baseline information on performance (not load testing, just basic tidbits) that will be useful as we go forward. It's quickly heading toward my "how I'll automate the deployment smoke tests" setup. I wish there was the same controls for Firefox, but I'll take what I can get - and the fact that it drives the browser is 100% benefit for me. The tests I develop using Watir will be very interactive in use.

I was mildly impressed with canoo web test. It's an excellent idea, but the only massive win that I can see that it gives me is an efficient way to store test sequences in XML and work those same test sequences through CruiseControl. Canoo seems to be moving actively towards providing a full DOM that I can explore and twiddle with testing, but it's just not quite there yet, and the javascript functionality with the 1.7 release was just quirky enough with our site to be really annoying. I guess I'd been hoping for more features and detail in Canoo - performance numbers and built-in bits that were farther along than the current Ant task definitions that are recorded. Of course, I can add my own (that being the job of open source).

Where Canoo was a nailed down black-box functional suite, Grinder is the load tester. Everything with Grinder is a python script running from within java and driving the Apache httpclient code to make things go. There's a lot more support than I explored in Grinder, so that's going to get more examination in the future. SOAP support, recording, etc - should give me the ability to quickly put together some good load test kits.

And finally at the lowest levels, HttpUnit does exactly what you'd think it does. All the detail, all the mess. The obvious downside being the relatively high amount of work it takes to develop the test scripts, the upside being you've got complete control. htmlunit got a looksee too - simply because it's clearly getting some attention from other projects and has the nice detail of encapsulating the response DOM's and all that good stuff already. Wrap the output with NekoHTML or Tidy, and you've got a lot of extra testing happening there. Canoo seems to be agressively moving to htmlunit, so it'll get more attention down the road as well - although for now I'll simply be starting out with HttpUnit just to get things cookin. Oh - and HttpUnit ties beautifully into Ant, JUnit, and CruiseControl as well - so we have quick integration capability right there.

I'd like to find something where I can use the HttpUnit or Canoo for basic load/performance testing. Once the test scripts/sequences are developed, it will be rework to migrate the same sequences into python so that I can run it through Grinder. There may not be any good way around that one though.

Now I'm starting to think about where I can store the long term trends of the testing - how long sequence X, Y and Z took - so we can see progress as we continue to develop. And most importantly, getting a baseline performance number nailed down for a number of user scenarios.

testing at a different layer - the browser

Well, it's past noon, which is the time I decided that I'd haul my butt down to a coffee shop and start digging back into some of the low level testing components (JUnit, HttpUnit, HtmlUnit, etc) so that I could catch up on all the innovations and changes that have been going on for the past year while I was out fiddling in other technology areas.

Before I headed out though, I decided to do some looking around and research to find out what else is out there at a layer or two higher. Yes, I was even considering the monster pay-for tools - all because what I was looking for was something that would give me reliable timing information for the end user experience in a browser. The low level tests are great, awesome, etc, etc - and they don't fit one of the needs that I have.

It turns out that the web application I'm aiming to test isn't well suited working at the http request/response level of components because there's large numbers of images that are dynamically generated and the implementation of how they load is a critical aspect of the web application I need to test. I first thought I'd look at some of the record and playback tools, but I inherently don't care for those because of the extreme fragility of those tests.

But while I was looking, I think I hit the jackpot: WTR. There's a similar game in Perl called Samie. They both drive IE through the COM interface in Windows. Ah, yes - just about exactly what I was looking for. And best of all, they have a DOM interface to the results, so I can query and poke and make sure all the bits came back where I expected them.

While I was digging around, I found a great reference to using WATIR (or WTR) at InformIT which even included sample code. Damn near a perfect "how-to" article for this technology. But I think the real win, which I've been reading through voraciously this morning, was Brett Petichord's publications page, which included a really cogent presentation on homebrew vs. commercial testing called Homebrew Test Automation (PDF)

Now I'm not sure if I'll get into Samie or WATIR, but I'm glad to have found both. It was a complete side track from what I intended to be getting into today, but way worthwhile. And I still have plenty of time to head to the coffee shop, drain a latte or two, and dive deep into the programming pieces again. Shoot - the biggest thing I think I'll be spending my time with is learning the new IntelliJ IDEA 4.5 IDE.

Now with all that good IE automation fun, I'm going to have to figure out how I can drive Safari the same way. WebKit gives some of what I'd like to see, but really I need decent access to the DOM too. It may be that some combination of PyObjC and WebKit is the way to go. Maybe it's reaching over the whole thing with Applescript - scripting from the upper level similiar to Samie or WATIR. And there's always RubyCocoa if I take this chance to learn Ruby, then I've a lovely tie in to Objective-C there.

hmmm. That whole WebKit/DOM thing - I think I'll need Tiger first.

18 days

18 days is the age of this lump of fur that we saw today at the Seattle Aquarium.

We were looking for something to do today that didn't involve a whole lot of computers for me (I'm thinking about spending most of tomorrow submerged in programming to get back up to speed on a few programming 'topics' and using a Java oriented IDE). Karen didn't see anything interesting playing at the Science Center IMAX, but noted that the aquarium had a new sea otter - a very young pup. I guess it's common for first pups (of which this is one) to not make it in the wild, but the aquarium was clearly excited and hoping for the best for this little fella. As we watched, it lurched itself into the water to attempt to paddle around, which it was able to do for all of 3 seconds before "mama" came and picked it and started doing circles with it.

Mostly the aquarium was packed, which isn't my favorite, but we found a few areas that were pretty open to watch the sea life wander around. Seattle has just a wonderful aquarium.

Scotch Ale

Not like I needed any alcohol, since yesterday was St. Patty's day, but I had a really lovely Pike Place Kilt Lifter Scotch Ale at the Hilltop Ale House this evening, and clearly needed to tell the world that it was tasty. That and a nice large bowl of gumbo made a late dinner for the evening after wandering around through downtown and basically window shopping.

St. Patty's Day

You can definitely tell it's St. Patty's day.

Our office is across the street from Fado - an Irish pub at 1st and Columbia. Since about 10am, there have been piper's playing continuously and there's currently a line waiting to get into the Irish pub.

This has got to be the best day for selling beer of the year.

The piping's been kind of fun to listen to, actually. I recognized the standards (Scotland the Brave, Amazing Grace) and I've been really enjoying the bits I don't hear normally. Not too bad having a live band outside your window, huh?

Oven in action

That new oven we got a month or so ago? Well, here it is in action!

Along with a very floured countertop. We made Teisen for dinner tonight. Yeah, I know - it's not very Irish, but we're having our Irish dinner tomorrow evening. Tonight it's welsh - Teisen Flats are welsh griddle cakes, and very yummy.

In welsh, at least according to my handy-dandy lexicon, teisen means "cakes".

Why does water do this?

I was impatiently waiting at a bus stop today and since I had nothing better to do, I watched the water flow over the pavement. As I was watching, the water approached a stain on the pavement, and I thought "Why is it avoiding that space?"

Maybe it's raised just slightly? Is it just the repulsion of oil and water doing something funny with surface tension? Something else I haven't thought of?

eclipse 3.0

I haven't been fiddling in java-land for a while, so it was only in the past few nights that I've pulled down the latest Eclipse IDE to see the state of the union.

Wow! Has that code progressed!

It's a little compact on the iBook's screen, but the overall effectiveness of the IDE was pretty darn good. There's even a subversion plugin (windows only darnit) for it.

I'm really curious where IntelliJ IDEA is these days. From their website they're at version 4.5, and I last saw them back to 2.7...

On the phone all day...

Sort of feels like I'll be on the phone all day today.

As well as posting job listings, I've also posted an apartment for rent on Craigslist. The response has been tremendous today - I just wish I'd planned this whole thing better so that I could have showed it some today. I've received seven calls so far, and 2 emails about it - so the response has been wonderful.

Pointy Stick Compendium Project

Just for john - the The Pointy Stick Compendium Project

job listing flamemail

I posted a job listing yesterday for DocuSign - we're hunting a test automation engineer to come join us. And it apparently sparked somebody's funny bone, because I received flame mail at work that looks surprisingly like this other posting on Craigslist: Testing is not automation.

I wrote the individual back, just to see what he'd say if there was a conversation involved, but haven't heard anything.

First time I've ever received a flame mail for a job posting...

Colloquy and Trac

I didn't realize that the Colloquy project was using Trac.. how very interesting! I'll have to dig around in there a bit to see how they've integrated it around, but from the outside at a glance, it's very nicely done.

iRooster icon

Hey cool!

Aaron's iRooster has a new icon, cleverly stuffed and hidden away in a nice review of the app

I like the icon Aaron!

flowering fruit trees

Karen and I just wandered back from a walk around Queen Anne that resulted in me eating way more than I intended, but at least got Karen some dinner. Tonight is a delightful night to be out walking - the scent of all the flowering fruit trees is all through-out the air, the air is just slightly cool, and somewhat still. So that lovely aroma just sits there, hanging as you walk through clouds of it layered under the white or pink blossoms which are slightly glowing with reflected light.

I love springs here.

UnitKit v 2.0

UnitKit version 2.0 has hit the streets! You can get the DMG online, peruse the 2.0 docs or if you've into a little subversion, get the goods through SVN:

svn co https://svn.x180.net/public/unitkit/trunk UnitKit

The new componentry using Notifications for it's messaging around the system, and from there it's not too surprising that Growl support got plopped in. Actually, I think the growl code in UKBasicReporter.m is a lovely example of 30 seconds to supporting Growl in your application.

Just remember - 2.0 is a work in progress. That's showing a bit in the docs (which I should probably help with, but I'm writing this right now instead - and besides, James is very good at writing docs...)

Microsoft and Groove

The word is out that Microsoft is aquiring Groove Networks, and all I can say is "Wonderful!".

Sharepoint was, and still is, a disaster of a product. Groove has been an "add-on" to Sharepoint, but my fervent desire is that we'll be seeing Groove as a replacement to it's Microsoft swag at collaboritive tools.

I haven't actually dug around in groove for over a year now, but I was impressed with their collabortion toolset. Albiet, all windows. Still, I think they did a damn fine job. Congrats on the aquisition! You played it out well.

ps: yes, I edited this one heavily. When I re-read the stuff above, I noticed that I did nothing but label interesting and varied profanity for almost all of the 2nd paragraph.

test case management software

I've been spending some of each of the past few days installing various test case management tools to get a good overview of what's out there today. There is a fantastic resource out there currently: Open Source Testing.

So from that list, I looked at web sites, perused sourceforge, and the following made my cut to try and install to see what they were about:

Bugzilla Test Runner, Fitnesse, QATraq, Testitool, TestLink, and TestMaster.

So the first one to die off was Fitnesse. Not because it wasn't interesting, cool, or any of that - but simply because it wasn't a test case management tool so much as an interesting cross between a Wiki and automated invocation. Or maybe I'm mischaracterizing it, but suffice to say that it wasn't structured enough for my needs and it looked like it had a reasonably high learning curve that I didn't want to incurr.

So on to the installs:

Bugzilla Test Runner is an interesting bolt-on to the Bugzilla that is a complete pain in the ass to install on a Debian SARGE box when you've used apt-get to install Bugzilla. The bolt-on scripts are reasonably well done, and the instructions good - based on the expectation that you installed Bugzilla from source and didn't do any directory layout tweaking - which Debian has had a good time with. So now - three days later - it's only partially installed.

QATraq and TestLink were gems to set up. I'd previously installed MySQL (due to the Bugzillar) and phpMyAdmin, so installing those two was primarily glancing at the docs, creating the correct databases, and using the provided SQL dumps to generate up the baseline information. Both are based on php & MySQL.

I have used TestLink at a previous job, and I was pleasantly surprised to see that it's undergone some updates and upgrades in just the past few months. The UI looks like it might be a LOT easier to use. TestLink maintains a demo site (maintain being very loosely defined), which you can use to see what the UI looks like. It was a concept of Products and TestPlans. Products have 1 or more components, and each component has one or more categories. Multiple products can be pulled into a single test plan, and there's nice features for importing and updating the test cases as you go along. There is some loose integration with Bugzilla embedded, and some noises about supporting the Mantis bug tracking system.

QATraq is the new kid for me. Quick to install, clearly a fair amount of work and thought put into the system, the first thing that comes to mind is that maybe the developers of this system are spending a bit too much time focusing on the overall bug process instead of making the test case management tool really optimized to usage. That being said, it's clearly a work in progress with loads more intended. QATraq also maintains a demo site, and I recommend taking a look simply to see the level of effort put into laying out the site. It's not my favorite design ever, but it's clear they put a lot of work and thought into making and polishing the app. While testlink has a really reasonable workflow for walking through and operating through the test cases and plans, QATraq is a little clumsy in that regard. Or more specifically, I haven't yet really grokked how to make it work efficiently.

Now neither of these guys yet support anything automated writing into the test cases... i.e. automated testing incrementing a "build number", running a set of baseline tests, and submitting the results. That's actually where I'm intending to go. With TestLink, we'd hacked in a component to enable us to do just that - submit test results via an HTTP request (some tests were running from a shell command-line only where the best we could get was a wget command to return results). That quickly led to checking out TestMaster.

TestMaster wasn't too bad to install, although I had to make a left turn on Debian to get some of the components on the system (Postgres, some Perl libraries, etc). One of it's most compelling features is that it already has a built in SOAP server-api so that automated systems can report into it. Which is a good thing, as the UI is so bare bones as to be almost hideously ugly. Just like something I think I'd make if I were making it. TestMaster has a Department -> Project -> TestSuite hierarchy and does a reasonable job of reporting out stats as well. I haven't yet sat down and tried to work through some test cases to see how well it works for manual setups, but it's certainly appealing from just having the API.

What? Yeah, I skipped over TestiTool. I should have paid more attention and noticed the development appears to have ceased in 2003, because this one was a complete pain in the ass to install. I'm sure the guys who are actively making it/working on it are in heaven - as it looks pretty nice. For now, however, I've bagged the whole kit because I certainly didn't want to spend 3 hours reading through code to understand what needed to be installed, how, and where. The hierarchy of Testitool is simply TestPlans contain test cases and you're off to the races. I applaud the relative simplicity, just not the install.

So now I'm down to reviewing my newly installed versions of TestLink, QATraq, and TestMaster. What I'm looking for, and my "review" of the three, will have to wait for another blog message - and for me to do the work.

one dude and his computer

One dude and the computer he uses at his office job is making the news. And of course, here I am helping to spread it. Not that it needs any help, given the wide publicity that it's already received.

The dude is Linus Thorvalds, primary instigator behind the linux kernel. And yeah, he's got linux on his dual G5 powermac. The funniest thing is that he apparently got the machine for free.

So what kind of news is that?

Your own google map

I don't know if this was ever considered or intended, but the ramifications for something like the wonderful tutorial at Engadget on creating your own "custom" Google maps is really quite large.

For one, all that information is in a well defined format that Google obviously knows quite well - as they more or less specified it. And now it's on the web - with links to additional content with lovely direct location ties. And now, if you want to use the url, you have to make a request to Google with the XML file you create as an option to that request. Talk about a boon for Google! If this sort of thing gets widespread, they've got a HUGE lead on really nailing down location based content.

MacTech

Somewhere over the past week, I received my copy of the latest edition of MacTech. The magazine has really undergone some changes - branching out quite a bit to reach a wider audience. I'm not sure I really like it, but I'm still reading it. I guess time will tell.

But one thing they had in this past issue was a reasonably coherent article outlining the differences between Postgres and MySQL. Now I've always been a MySQL fan, but that was mostly out of default. I never really investigated Postgres, although I've been prompted to do so by some of it's fans. The article makes some claims that it doesn't really back up with numbers, but even still it's pretty good.

wonderfully disconnected

I've been wonderfully disconnected this weekend, chilling out at home and not doing much of consequence. Well, that's not true actually - I've been plenty of consequence for me, just not anything for anyone else. I spent time mowing the yard (well needed), cleaning house, watching movies, and generally being a complete veg on my first full weekend off in a month. And damn, it was needed.

This past wednesday our family gained my first neice - Emily - and reports are the everything is going well there. I got a picture of her via email today. Blonde hair and blue eyes - very much like her father. But that's just at birth - we'll she how she grows. I expect the blonde won't hold out.

bad bus karma

I'm glad today is wearing to an end. It's been a bit of a rough one.

I've been sleepy and tired all day, the dragging of which gradually resolved into a headache that's been plaguing me since about 1pm. I spent this evening taking an easy route home, stopping at a bookstore to snag a couple of technical books I've been meaning to get for the past week, and then catching a route 4 up the hill.

I guess it shows how just darn grumpy I am today, as the bus driver completely ignored the southern route 4 loop this evening. I was thinking all sorts of vile thoughts, when all I said was "isn't the route four supposed to lay over at Nob Hill and Galer?". The response I got was "oh, I made a mistake. I wait here."

Great. Jerk. If I'd wanted to walk home from Boston and Queen Anne, I could have done so without touring all of the route 3 loop.

At least I've got some more advil in my system now, and the cats are fed, so I suppose I'll just try and take it easy for a while now and loose some of this grumpy and headache.

But you've got to admit - that's definitely some poor bus karma to pull that one.

It is wrong

It is wrong to think that a machine should require a reboot to function properly and normally. Yet that's where I am with Windows XP Professional.

Just had to spew that.

Trac

Have I announced to world that it's really nice to be able to quickly and easily edit web applications that are based on Python? Ah, it's SO nice.

I just finished installing trac today, fiddling with using it in a project. It includes a ticket tracking system, and like every other ticket tracking system it has a pile of severity and priority enumerations.

I understand that most people like a variety there, but to me it's the worst of all possible setups. You get 8 levels of priority, and people end up spending hours screwing around with arguing is the priority (or severity) level 3 or level 4. All this crap. I have since switched to the minimalist approach to getting severities and priorities. You get three leves - Top, Middle, Bottom. That's it.

The problem comes in with the software. Bugzilla, Trac, you name it - they all seem to include at least 5 priorities or severities. So being able to reach in, and tear out a few levels is important to me. And with Trac, it only took a few seconds.

I started looking at the code, and in about 2 minutes I'd tracked it down to a database default setting and edited the file db_defaults.py. Bam - now we've got three priorities and three severities. I started to do the same thing in Bugzilla. Ug. I didn't make it past the 10 minute mark, when I decided that I'd have to spend some more serious time understanding the code before I reached in and started tweaking.

Now I'm sure that has to do with how bugzilla is programmed, but it's still darned nice.

Constantine

I decided I needed to get out and do something tonight, so I took in a late show of Constantine at the Cinerama. And I'm really glad I did. I'm not super excited about Keanu Reeves, but they keep casting him in movies in which I'm interested - often as leading characters. Oh well. The character that I really enjoyed was played by Tilda Swinton - who did a lovely job at the androgynous and somewhat pissy angel Gabriel (although I thought Peter Stomare did a good job as Satan too).

Now really, it's nothing like the DC Vertigo comic - well, almost nothing like it - but it was an enjoyable romp. The special effects were suitably impressive, and I thought their rendition of hell was pretty darn good.

Alone for a few days

Karen's off north to Oak Harbor, so I'm on my own few the next few days.

And to start it off, I have my internal clock completely and total screwed up. I was up really late night - like 3:30am, and I've been pushing my hours quite a bit of late. And then, the critical day (ngiht) at the office was complete, so I slept in. And I really slept in.... I walked into work today at 3pm to a lot of smirks.

Yeah, that's getting to work "a bit late", isn't it.

I actually expect that I'll be able to sync my internal clock back around without too much trouble, but it'll be interesting to see.

warm evening

It's a fairly warm evening out tonight. Warm being 58 - that "just perfectly comfy wearing a sweatshirt" temperature (YMMV). The night's clear - all very different from what I expected. The weather looked like it would be rain all day - at least from the satellite. And it was this morning. But this evening cleared up, was even quite warm riding the bus home.

3/1

heh, Happy 3/1