Older blog entries for purcell (starting at number 29)

A rainy public-holiday Monday in Bavaria...

I'm gradually getting around to acting on the patches and suggestions people have sent me over the last few months. Today it was the turn of Jargs to get a facelift, a polishing and a minor release.

After months of concentrating on photography, paid work and unstructured PHP hacking, perhaps I'll manage to release some more useful things. But only if the weather's not 29 degrees and sunny, because Biergärten and laptops don't mix!

Spring has arrived early in Munich. While hibernating I've had almost nothing to do with coding, and I've failed to reply to so many e-mails that I should probably be excommunicated from the open source world.

My penance has begun with an effort to catch up on those mails and patches. I've just made release 0.4 of WebUnit which adds HTTPS support courtesy of a patch from Oliver Rutherfurd. (That's the first release since last July, I think.)

If anybody reading this sent me mail, I apologise; you almost certainly haven't had a reply yet. I could claim a disk crash, but that would be a disservice to my ThinkPad.

Release numbers are a strange thing. It seems apparent that there is no logical scheme for numbering releases over the lifetime of a piece of software, other than that release numbers should usually increase as time goes on.

For some programmers, I imagine that release numbers are proportional to the length of the feature list, the number of lines of code, or even the number of bugs.

Not for me. For the record, my release numbers are not particularly intended to converge on 1.0. What would 1.0 mean anyway? Instead, I earn release number increments. Usually I only earn '0.1' at a time, but sometimes I feel like I've earned more, so I bump the release number by 0.5.

Works For Me™

Version 0.6 of DBDoc is now out, featuring Oracle support! (For the uninitiated, this means that javadoc-style HTML schema documentation can now be automatically generated from a running Oracle instance.)

(Most of the hard work was done by Andy Todd, to whom I am greatly indebted.)

This release neatly coincides with the change from PostgreSQL to Oracle that has taken place in my day-job this week. The wonders of synchroncity...

Having used Oracle for 2 1/2 years on a previous project I wasn't looking forward to returning to its quirks and complications, but the experience has been quite bearable so far.

Andy Todd contributed a fantastic Oracle implementation of the DBDoc schema API. It's now in CVS and is receiving some last-minute tweaking before a new release is made. Watch this space...

Last night it snowed enthusiastically. Accordingly, I have another recent little poem to offer:

Cold air shrinks my brain
No space left for worries.

A quiet weekend at home; hacked up a Python module that parses Java class files. It was mainly just a fun exercise, but it could be entertaining to implement the Java reflection API in Python; it could then be integrated into non-Java IDEs, for example.

And, completely off-topic, a fun poem I wrote last week

Winter day in Autumn
No-one wearing gloves but
I'm glad I brought mine.

Today the first snow of the winter is falling here in Munich! It's such a joy to watch the seasons change. It seems like only yesterday that it was 35C and everybody was swimming in the lakes; now's it already creamy soup weather, and soon we'll be ice skating!

Is that relevant to the world of software? It is, but making the connection is left as an exercise for the reader.

Wow! dbdoc was pretty well received; 800 page views and 30 downloads yesterday. Someone kindly volunteered to implement the schema API for Oracle (and possibly MySQL too).

And someone else pointed out a couple of shortcomings in my simplistic initial version of the API. To my mind, this is one of the nicest things about the open source community; if one gives away vaguely useful software for free, other people will freely offer their hard-earned wisdom in return.

Time to find out some more about the quirks of various databases...

After a long coding lull, I've just made the fourth release of 'dbdoc', and announced it on the Python list in order to solicit implementations of the schema API for other databases.

Day-job commitments, good weather and social activities (shock, horror!) have kept me too busy to hack on free code in recent times. One highlight this month was meeting illustrious Erlang hacker Luke Gorrie here in Munich at Oktoberfest time. (Another was seeing the post-dotcom-crash industry decimation at the Systems IT show.)

I posted my DB schema documentation generator at dbdoc.sourceforge.net.

It works pretty nicely for me, with Postgres 7. Will anyone volunteer to implement the schema API for another database?

At the weekend I hacked up an initial version of the DB schema documentation generator I mentioned in my last diary entry, for Postgres 7.

It's about 400 lines of code in total, and does a pretty nice job for the schema we're using at work. Auto-documented are: tables, columns (types, lengths, default values, nullability, primary/foreign keyness), and indexes. Stored procs should be easy to add, though we're not using any because they're a Bad Thing For Portability.

I'll tidy up the code and post it somewhere; then I'll be looking for experts to hack up implementations of the schema introspection API for other databases.