|Home » Resources » Rants
Bloat Dissections II
Week of June 8, 2000
(Edited from the RISKS article published simultaneously)
The bloat dissections continue. One remaining question in the RISKS article of 30 April 1999 and its follow ups is: 'would dissecting such an application as RegClean at source level really reduce its bloat?'
Curious coincidence then, that I should happen upon the source code to another Microsoft Registry cleaner - RegMaid, from 1993 - 95 (btw, this application might very well be the basis of the later RegClean project as well, judging from the embedded resources).
- Fact: the distributed executable to RegMaid is 153,600 bytes.
- Fact: it is distributed with its source code.
- Fact: after less than ten minutes work I got it down to 69,632 bytes.
This is a savings of 83,968 bytes, or well over half the original image size. This is a new image only 45% the size of the original. And in less than ten minutes. I got 83,968 bytes of bloat off my disk that fast.
Not to beat the dead horse but - just think about it. Ten minutes, 55% savings of 83,968 - and I had never seen this code before in my life.
If this doesn't prove that bloat is inexcusable, I'll just have to send more examples. Cutting the bloat in that application, and attempting to estimate the cost of doing so, really misses the point too: the code itself is rather intricate, and this Rome was not built in a day. Barring the author's getting the architecture completely right from the beginning, the ten minutes it took to shave the EXE by 55% represent:
- Half of only one coffee break.
- A late arrival into the Thursday night Quake game.
- One final tweak in the evening before dinner is really served.
- Something to do on the laptop while commuting to work in the morning.
It's trivial. Avoiding bloat has never been an effort, despite what the defenders of latter-day commercial software like to claim. Things can be done right from the beginning, or even if not, corrected in a negligible envelope of time.
As for really astounding comparisons, try the following - where we got a monster down from 3.5MB to just 7KB (seven kilobytes) in a single hour.
It's professional pride on the one side - and 'who cares?' on the other.
(Rick for radsoft.net, RISKS Digest June 8, 2000)