> It's interesting to note that the 200MB used by Office 2000 costs around
> $3. How much would 5-10MB of disk space have cost in 1988? I bet more
> than $3. So what is the proper reaction to bloatware? Is it to be
> morally outraged because "in my day, we fit programs into 1MB, and we
> liked it?" Rather, as somebody on Slashdot pointed out, we should
I'm sorry, but this is a typical Microsoftianism. Code bloat makes
itself felt when your app doesn't fit into the core, and your cache
runneth over, miserably. Spinning bits are sure cheap, but that's
not the point.
But most horribly, code bloat corrupts hardware. It would be dead easy
going wafer-scale integration with stack processors using 1 MBit
embedded RAM technology years ago. A MISC core with networking is
what? 12 kTransistors? Sorry, can't do that with Pentiums.
A Forth OS can be as small as 2 k, and thus occupy a small fraction
of a MBit core grain. Don't try this with anything else than Taos, L4
or QNX, forks. And even if your OS is nanokernel it doesn't help if your app
can't be decomposed into a sea of tiny objects living on the wafer.
See any way of turning Office 2000, or gcc, for that matter, into that
rich object soup?
Please don't take it personally, but I think we should have issued a
preemptive nuclear strike against Redmond a decade ago. Of course with
even some more foresight we would have fused Intel and Big Blue to multihued
radioactive glass in mid-70s.
> separate size bloat from feature bloat. People who complain about size
> bloat should join the 90s and buy a new hard disk. Feature bloat, on the
> other hand, is actually harmful since it makes these products a severe
> pain to learn and use. There are so many menu options in Office 2000 that
> by default it hides infrequently-used menu items. Scary.
Ever seen StarOffice under Linux? Horrible. Just horrible. "Do
everything in one place" is not a yota different to "where do you
want to go today".