You may not, but others might. Or maybe the results would show that most people don't, who knows. Without a side by side, apples to apples comparison it's really hard to know whether someone needs commercial software tools or not.

I run Linux on my desktop and only run WinXP within VMWare for the very occasional Windows application I need to write or maintain. Every piece of software on my Linux desktop was free, in fact included in the Linux distribution itself (except VMWare, funny how you have to start paying as soon as Windows gets near anything). I have nothing against free software whatsoever--some of the best software is free (Apache web server, PHP, MySQL, etc.).

But my desktop has 100GB of hard drive space, 768MB of RAM, and runs at something like 1.6 GHz. Getting every last ounce of performance out of it isn't particularly critical (especially when getting every last ounce of performance out of my HP laptop causes it to overheat and shut down immediately with extreme prejudice).

Only with a fair apples to apples benchmark will we truly know whether commercial compilers for the 8051 generate sufficiently more efficient code (in terms of speed and size) as compared to free software on a platform where the difference betweem 8200 bytes and 8192 bytes could mean the difference between going to the next higher (and more expensive) derivative option. The results might show that free software is just as good as commercial software; or it might show that it isn't. Conventional knowledge says the commercial compilers are significantly better. It'd be useful to test that conventional knowledge.

Regards,
Craig Steiner