Mandala is a program that I wrote in the 1990s to generate eye-candy. It also runs as a windows screensaver. It has a home again at http://www.frogmort.com/mandala/
This is because a week or two ago, a cryptic message appeared on my LJ

(Anonymous) (144.5.140.22) wrote:
Sep. 17th, 2008 10:40 pm (local)
Mandala
Do you still have Mandala for download on the internet?

So I uploaded a zip file and replied, and got a post on another random LJ entry a few days later:

Mandala
Thanks so much for the zip file, cant live without that screensaver!

The IP address appears to be in the Wichita area.
The reminded me that it wasn't on the internet. So I have uploaded it, with docs.

...is a worthy goal. Given that LJ's unique selling point is the friends lists, friends-locked posts and all that stuff, The Xhtml Friends Network looks like an interesting stab at the problem.

But how would I go about, given someone's name/url, find out who their FOAFs are? How would I retrieve posts that are locked to friends only, given that I am on a friends list? This would need authentication, yes? Would it be centralised?

Wait, it doesn't do any of that at all? Rats. And I have to go back and edit all old posts when I change friend status? The information isn't stored in one place? Oh well, is anyone else taking another stab at it?

Obscure code behind the cut

( Read more... )

Delphi news for all zero of you who care:

Highlights:

* Next release is aimed for the first quarter of 2007. Main goal is .NET 2.0 support and first things for Windows Vista.
* New .NET language features will be ported to the Win32 compiler. They have mentioned Generics

We know that in the greater scheme of things this is all old hat, but not for those still maintaining the code.

I reread the Joel test at regular intervals.

And though I agree with more than half of it, there are inclusions and omissions that puzzle me. These tests are relative to current industry practices, e.g. nobody puts "do your employees have chairs, or do they have to stand" on this, since everybody does that. I've tried to remove practices that are too narrowly applicable, or are just one possible way of meeting an end.

So here's my tweaked version - The Frog test: 12 steps to better code, shamelessly copied from Joel. The ones that differ from Joel are marked with a * . I've reordered them to put similar ones together.

( Read more... )


My current employer scores around 5 (half marks for 2 and 10: we sort-of have a bugs database, and it's usually quiet enough.) which is not as bad as I thought. But not good anyway.

So what do you score?

Signs You're a Crappy Programmer (and don't know it)

He also dislikes the pointless ban on multiple function exit points

Earlier this week I drew up a list of what I wanted from an ideal revision control system. Reading it a day or two later, it looked like basically subversion with a little perforce added.

So today I got onto subversion. Once you get the hang of it, creating and using a SVN repository is preposterously easy. There's even extra feaures that I never thought of, but I can see the use of them. E.G. the "blame" feaure will give you for each line in a file, the person who last altered it. Heh.

So I made a repository for Mandala on my machine's E: drive, and checked it out to C:Code\Mandala_SVN\. Revision tracking and a backup on a second hard drive: Excellent. Oh, and I made a new version of Mandala.. The bitmap dump works and the spirals are a bit more varied.

---

Dr Who was a great story, with the hot and brainy Madame de Pompadour haunted! By robots! Somewhat sad on the transience of time, especially when the people you know use short cuts through it. But Dr. Who seems to stick to orthodox chronology, they avoid the loops like "this is the first time that you've met me, but the third time that I've met you" or "you spoke to me last March? Oh well, then, I'd better go back and chat to you then."

Reading Jeanne-Antoinette Poisson's bio on Wikipedia again, it seems that they glossed over a lot. Oh well.

Public int IsThisSoBad(int param)
{
   If (param == 0)
     Return 0;

  .... more code ... 
}

Yesterday I finally got an idea as to why some people seem to think that it's a "law" that a function should have only one exit point (one return statement, if you're in Java/C#/etc). It happened in This discussion on The Daily WTF

( Read more )