> >>>>> "mh" == Mark Hahn <hahn@neurocog.lrdc.pitt.edu> writes:
>
> >> I think a reduction of over 20% and close to 1.1 MB is a major
>
> This assumes that we all have T1 links. It took me about an hour and a
> half to download the new source the other night.
The point about time to download is valid, just relatively unimportant.
One approach would be for the mirror sites to run the source file through
split. How about YOU have a look at the mirror program in use at a
nearby mirror site, work out and test a configuration that will
automatically do the split, and persuade the site administrator to use it?
You would still download the entire kernel source, but in sessions of
convenient length.
What has annoyed some people is that you want someone else, someone at the
centre (specifically, Linus), to do a lot of work and make a fundamental
change in the distribution for a totally peripheral purpose touching on
your convenience. The sensible course, on the contrary, would be to keep
out of that person's way and not get under his feet while he is very busy
with work of much greater value of which we, if we allow him to do it,
will be the beneficiaries.
It is arguably worth dicing the kernel well downstream from the source.
This has been mentioned as an option by several respondents to the
proposal, but they have been ignored and we return incessantly to the
request that someone else to do it at source. Okay. We heard you the
first time. Consider the suggestion to have been well and truly made,
and let us pass on to something else.