Xref: utzoo comp.sources.wanted:16351 comp.text.tex:6916
Path: utzoo!utgpu!watserv1!watmath!att!emory!wuarchive!zaphod.mps.ohio-state.edu!rpi!batcomputer!cornell!uw-beaver!zephyr.ens.tek.com!tektronix!percy!m2xenix!quagga!hippo!spel
From: spel@hippo.ru.ac.za (Dr. E.W. Lisse)
Newsgroups: comp.sources.wanted,comp.text.tex
Subject: Re: sorter for large ascii files for msdos
Message-ID: <spel.671542996@hippo>
Date: 13 Apr 91 11:43:16 GMT
References: <spel.671470036@hippo>
Sender: usenet@quagga.ru.ac.za (Rhodes University NNTP server)
Organization: Rhodes University, Grahamstown, South Africa
Lines: 65

In <spel.671470036@hippo> spel@hippo.ru.ac.za (Dr. E.W. Lisse) writes:

>Hi,

>I need a SORT program that can handle 20000 lines ascii and is
>reasonably fast. I do not have ftp access nor can I receive
>comp.ibm.binary or what it is called as we have a phone link to the us
>and it has too much traffic for the finance section :-)-O

I today discovered I need one that can handle 50.000 lines. (see below)

>I used Timo Salmi's TSCHEK recently and it is very nice. I read in the
>commands from LATEX.TEX deleted the internal commands (\??@????) and
>added them to the word list. Unfortunately I have to resort to very ugly
>tricks to get the stuff sorted fine. (like reading it into data bases
>and stuff :-)-O)

I dug around in my ZIPs and found that other nice speller MicroSpell
from MicroEmacs. I reread the docs and found a way of decompressing the
dictionary to ASCII (I always maintained it saves to read the manuals
:-)-O) Now I have merged the two dictionaries (13.000 and 48.000 words,
400KB)

>my sorter does unfortunately sort like this

>things
>thing

>instead of

>thing
>things

>It is however quite fast. 56 seconds for sortinmg through 13218 records
>(127288Bytes) 232618 comparisons and at least half of the time
>reading/writing to/from the 20ms hard disk.

It barfs now on 61.000 words. (This is forgivable :-)-O)

>SO, please get me the host and subdirectory and name of the ultimate
>sorter or email me the offer of emailing it to me. Sources would of
>course be best (so I learn some :-)-O) Tube-c and tube-pascal, but the
>binaries do it as well.

Yeah, give it to me :-)-O

I am willing to share the secret of how to uncompress the DICT.DCT for
those unwilling to read the manual :-)-O

>regards, el

>ps: followup can be directed to these newsgroups, as I DO read them
>:-)-O


>--
>Dr. Eberhard W. Lisse       (spel@hippo.ru.ac.ZA)
>Katatura State Hospital     (formerly extel@quagga.ru.ac.za)
>Private Bag 13215           (Real Soon Now ...  el@lisse.NA)
>Windhoek, Namibia           (no FTP yet. [This is Africa :-)-O])
--
Dr. Eberhard W. Lisse       (spel@hippo.ru.ac.ZA)
Katatura State Hospital     (formerly extel@quagga.ru.ac.za)
Private Bag 13215           (Real Soon Now ...  el@lisse.NA)
Windhoek, Namibia           (no FTP yet. [This is Africa :-)-O])