Blog entries from October, 2003

Today at work we had Damian Conway in to present his Understanding Regular Expressions course. Overall it was good even though we only covered half of the material. This was because Damian would often talk about fundamental aspects of Perl related to his examples. So while that detracted from the regular expressions it added a great deal more.

A some stage we need to do his Perl Best Practice course, assuming he has finished planning it…

The course also brought back memories of having Damian as my lecturer for a couple of subjects back in my early years of university. His Object Oriented Perl book also brings back memories of similar concepts, except in C++ back then. I must finish reading it a some stage…

One of my orders from Amazon arrived today:

• Extreme Programming Explained
• Extreme Programming Refactored
• Peopleware
• Code Complete

This time it came from Germany, last time I ordered from Amazon (Pragmatic Programmer, OO Perl) it came from Hong Kong.

Now all I have to do is find time to read them…

Actually old hardware, but new to me so it is still good…

I picked up a second hand box – Dell Optiplex GX110 small form factor, Pentium 3 667, 128MB, CD, floppy, onboard video sound network – at the swapmeet today in order to upgrade gromit.

Pity I won’t have the time to shuffle hard drives until next weekend…

I’m coming to the end of XP Explained and the car metaphor has been mentioned again. Basically the metaphor being used is that there are two ways to drive a car; point it down the road and then start it up, or start it up and continually steer so that it stays on the road.

Pragmatic Programmer talked about the same things but used a gun as a metaphor; to hit a target you can calculate the angle to aim the gun based on positions and environmental conditions and then fire a single bullet, or you can use tracer bullets in a machine gun and adjust your aim after you have started firing.

The thing that I keep thinking is that they are both talking about an old idea that even Mythical Man Month touched upon: Iterative Development.

This is again exposing the biggest issue I have with development strategies like XP (don’t get me started on ‘Agile’), someone has taken a variety of established (not necessarily utilized) ideas and wrapped them together as something new. The other issue is that because of this a lot of people now have an all or nothing mentality: “I don’t like XP so I will exclude all of the parts of XP as well.”

There is probably also a bit of fear of the unknown in there as well. The other day at work I sent around a link to WhyNotEnoughRefactoringHappens in response to an email. A comment was made about how it was good that there were all these related topics. However when I informed them that they were actually looking at a wiki their attitude instantly changed. To their defence they were introduced to wiki’s in the wrong way, basically “wiki is there, use it”. No introduction about what a wiki is, the benefits, and the disadvantages…

I’ve been reading through Kent Beck’s Extreme Programming Explained, an article with an example of refactoring perl, and more of the WikiWikiWeb. One point that they, and all the other stuff on refactoring I have read, say is that you must be able to test both pre and post refactor.

I can see how this is done for a module; write a test harness that tries various things and expects both successes and failures. But how do you write system test when the system is a web interface to these modules?

I have tried and failed but if I remember I will again try to search for some examples people have written, specifically with regard to testing web interfaces…

Last night I decided to bite the bullet and ordered a bunch of books from Amazon. Even with the postage to Australia it still works about about two thirds of the price they would have been to buy them at bookshops locally.

I have also just spent the last three hours reading through more of the WikiWikiWeb. I didn’t realise the true extent of the software engineering discussion on there.

Following on from my rediscovery of Wiki’s yesterday I have found a lot of pages that I want to read but just will never be able to find the time…

One page that I did manage to find time to read was How to Write for the Web by Jakob Nielsen. I regularily read his Alterbox column so I must have already read it but I can’t remember doing so…

Anyway, the point of bringing this up is that it confirms my philisophy that you should be lazy and say all you need to in the minimum amount of text possible.

At work today I was moving the style guide into the wiki which led me over to the original WikiWikiWeb to see if I could do an InterWiki link for Rob Pike (I included some quotes he made about software engineering in the style guide).

Anyway the point I am getting to is that it reminded me that I had been meaning to have a look at The Practice of Programming and it was interesting to see that a lot of other people share the view that all these ‘new’ software development techniques – ie Expert, Agile, Aspect, etc – are basically just a rewording of the practices that have been followed for some time.

I have also added Informal History Of Programming Ideas to my list of things to read…

Some other books that have been recommended to me are: * Peopleware * Code Complete

I spent some time last night and this afternoon stripping down and then rebuilding my media box with some changes:

• Replaced the aging 8GB boot drive with a new 40GB (now also a location for backups)
• Added in a third 80GB drive (rotated out of my main box)
• Rebuilt the LVM volume with the three drives and used XFS instead of ext3
• Moved over from Redhat 9 to give Mandrake 9.1 a try.
• and cleaned a lot of dust out…

Although Mandrake and Redhat are very similar there are a few quirks that I found surprising and will have to get used to…

Not much.

Something went seriously wrong with an upgrade to the network at work and all day today (possibly extending to tomorrow morning) nothing was accessable. Since we depend on the network, both development and writing documents, not much gone done.

Apart from moving my desk so people didn’t have to walk behind me to get to the shared printer I did manage to read through, thanks to a printout, a sizeable chunk of Joel on Software, in particular The Joel Test: 12 Steps to Better Code. What he is saying just makes sense, so why don’t we do most of it?