On Mon, 12 Aug 2002, Alan B. Pearce wrote:

>>My major gripes with MPLAB has always the 16-bit code, the clunky
>>simulator (although I do use it once in a while), the AWFUL support for
>>3rd party compilers -- especially source level debugging in C. Some may
>>say it's not realistic to ask Microchip to support 3rd party tools,
>>BUT...
>
>I cannot understand why they have not caused the Borland compiler they use,
>to produce 32 bit code. Perhaps Microchip have been developing MPLAB on Win
>3.1 machines :))

Thank the powers that be that they haven't because 16 bit MPLAB runs fine
under Linux/Wine by grace of this "fault" ... until the GNU tools will be
integrated properly <duck behind large flame shield>.

Peter

>So yes, there are aspects about it that are junk. It is not unreasonable to
>expect it to improve as new versions come out, not break existing bits -
>that is why people use HLL's, isn't it??? :))

Just for laughs compute the number of man-hours required to test every
feature wrt every other feature, assuming 1 hour per bugfix and an initial
'imperfection' of 0.1% (one bug per thousand features - don't forget to
count as features to be tested things like testing scrolling at each
corner, ltrb, and middle of window, maximized or not, with random data,
above or under windows etc etc). If you still think such a thing should be
free, say so ;-). F.ex. a program with 2 partially overlapping text
windows that do just scrolling of text (no editing, no nothing) will have
some 81 tests to pass, depending on how you count ...

Peter

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.