[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Debian Policy based on the wrong technical assumptions

>>>>> "Fernando" == Fernando  <alegre@saturn.superlink.net> writes:

    Fernando> I fear the Documentation Policy is being based on the
    Fernando> wrong technical assumptions.

    Fernando> Please take a moment and read the following timings made
    Fernando> in a 386 SX-25 with 5MB of RAM. It will help you
    Fernando> understand what is at stake. I did the test as fairly as
    Fernando> I could and I claim to be a honest person.

 I'm not seeing the dramatic differences in times that Fernando does.

 I just ran the same tests.  The subjective difference, here, is far
less than the empirical one... meaning that for the short time it
takes to format a man page, if Lynx is a second or two faster, then
what difference does it make?  On a 386-sx with `only' 5Mb of RAM, it
makes a lot more difference.  Waiting over a minute and a half for a
manual to format would be excruciating.

     (Fernando, buy some RAM.  You need it bad. :-) It makes a BIG
difference to go to 8-12Mb.)

bash-2.00$ /usr/bin/time lynx -dump \
> http://localhost/cgi-bin/man2html/usr/man/man1/telnet.1.gz >/dev/null
0.98user 0.76system 0:02.45elapsed 70%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (7126major+2355minor)pagefaults 0swaps

bash-2.00$ /usr/bin/time  man -l /usr/man/man1/telnet.1.gz >/dev/null
Reformatting telnet.1.gz, please wait...
3.58user 0.09system 0:03.95elapsed 92%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (1189major+500minor)pagefaults 0swaps

bash-2.00$ /usr/bin/time echo >/dev/null
0.00user 0.00system 0:00.07elapsed 0%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (72major+12minor)pagefaults 0swaps

bash-2.00$ /usr/bin/time snarf \
> http://localhost/cgi-bin/man2html/usr/man/man1/telnet.1.gz - >/dev/null
0.00user 0.02system 0:00.65elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (74major+16minor)pagefaults 0swaps

bash-2.00$ /usr/bin/time snarf \
> http://localhost/cgi-bin/man2html/usr/man/man1/telnet.1.gz - \
> > /tmp/telnet.html
0.01user 0.01system 0:00.64elapsed 3%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (74major+16minor)pagefaults 0swaps

bash-2.00$ ls -l /tmp/telnet.html
-rw-r--r--   1 karlheg  karlheg     32944 Jun 29 13:44 /tmp/telnet.html

bash-2.00$ /usr/bin/time lynx -dump /tmp/telnet.html >/dev/null
1.07user 0.67system 0:01.85elapsed 93%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (7115major+2349minor)pagefaults 0swaps

bash-2.00$ gzip -9 /tmp/telnet.html

bash-2.00$ ls -l /tmp/telnet.html.gz
-rw-r--r--   1 karlheg  karlheg      8576 Jun 29 13:44 /tmp/telnet.html.gz

bash-2.00$ /usr/bin/time lynx -dump /tmp/telnet.html.gz >/dev/null
1.18user 0.60system 0:01.87elapsed 95%CPU (0avgtext+0avgdata 0maxresident)k
0inputs+0outputs (7356major+2439minor)pagefaults 0swaps

bash-2.00$ uname -a
Linux bittersweet 2.1.36 #9 Sun Jun 22 15:16:13 PDT 1997 i586 unknown

bash-2.00$ free
             total       used       free     shared    buffers     cached
Mem:         63276      54024       9252      15464       2752      18456
-/+ buffers:            32816      30460
Swap:       122968       4940     118028

bash-2.00$ bogomips
Calibrating delay loop.. ok - 197.37 BogoMips

    Fernando> Please base policy on facts, not on pre-judgements.

 The fact is, bringing your system up to 8Mb of RAM would make a
*HUGE* difference in performance!  (as would a processor upgrade. Sock
away some cash, I guess.)  I found that the hop from 8Mb to 12Mb made
quite a lot of improvement.  The computer swapped a lot less, I was
able to run X windows and XEmacs then.  The jump from 12Mb to 16Mb
wasn't quite as dramatic, but did make a noticable difference.


 So, for the people with ancient 3 year to 5 year old computers, we'll
need to package docs in html.

mailto:karlheg+sig@inetarena.com (Karl M. Hegbloom)
Portland, OR  USA
Debian GNU 1.3  Linux 2.1.36 AMD K5 PR-133

TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .

Reply to: