Text 588, 207 rader
Skriven 2000-04-28 21:26:00 av Patrik Järnefelt (2:201/145)
Ärende: PK
==========
Den 14/4 dog Philip Katz, skapare av PKZIP. Han blev 37 år gammal.
Jag såg nyheten för några dagar sedan när jag skummade igenom
alt.folklore.computers eller någon liknande newsgrupp. En sväng över till
slashdot bekräftade det hela. Jag valde några trådar och läste lite om vad
folk diskuterat och en kommentar stod ut från mängden: "You're not mourning
a person, you're mourning a software." Det var mycket nostalgi och
tillbakablickar. Jag kände inte Phil Katz men jag har använt hans mjukvara
och använder den ibland även idag. Oregistrerat tyvärr.
För något år sedan funderade jag på att göra ett slags index för fidonews
där man kan följa saker som återkommit under åren, problem, utvecklingen av
mjukvara/hårdvara, det politiska läget. Det vore också intressant att följa
upp vad som hänt efteråt. Naturligtvis visade det sig att jag inte hade tid
med en 'annotated fidonews' men jag har iaf plöjt igenom samtliga nummer och
delvis strukturerat upp olika händelser. Komprimeringsfrågan är en av de
intressanta 'problemen' i fidonets historia. Så jag tänkte klippa och
klistra ganska mycket här och nu för att öka medvetenheten om nyttan i att
bevara nutiden. En gång kommer även vi att vara historia.
FidoNews 5-27, 4 Jul 1988
SEA vs PKWare -- What's It About?
by Ben Baker -- 44/76
I suppose most of you know by now know that System Enhance-
ment Associates (SEA) is suing Phil Katz and PKWare. There has
bee a lot of comment (I hesitate to call it discussion) on the
suit in conferences in FidoNet and the commercial services. Most
of what I have read gave me the impression that the writer
thought about it for at least 30 milliseconds, then flamed! So
what's it really about?
First, a little history. In CP/M days, there were a number
of utilities for compressing and decompressing files, based on
the Huffman coding technique. The most popular were called SQ
and USQ, but there were several variations. There was at least
one utility called LIB, which did not do compression (remember
the total memory space was only 64K, and CP/M used at least 4K of
that), but it did collect files, "squeezed" or otherwise, into a
single file so that they could be treated as a unit.
When MS-DOS came along, many of the old CP/M utilities were
ported. Among them were several (often incompatible) variations
of SQ/USQ. Our own Tom Jennings ported LIB to DOS. DOS BBS op-
erators then had all the functionality they had in CP/M; they
could compress files, and they could pack them into "library"
files, in separate steps of course. But DOS wasn't memory poor
like CP/M. The climate was right for something "new." Enter
Thom Henderson.
Henderson, one of SEA's principals, borrowing from concepts
developed by Brian Kernigan, wrote a "library" utility which
overcame a limitation of LIB by using a distributed directory in-
stead of a fixed-length directory at the front of the file. It
also had built-in Huffman code compression, eliminating the need
for SQ/USQ. He called it ARC. Almost overnight, it became a
standard among bulletin board operators.
As ARC developed, it acquired a number of useful features,
encryption and LZ compression, for example, stirring interest in
the commercial marketplace. Thus ARC became one of those prod-
ucts marketed both commercially and as shareware.
In an effort to encourage porting ARC to other systems such
as Unix, SEA made the sources for ARC available for download on
its bulletin board. These files bear the SEA copyright notice,
and before people may legally do anything with them other than
study them, they need SEA's permission. When someone asks per-
mission to port, it is granted with three restrictions on the re-
sulting program: it may not be sold, it may not be used commer-
cially and a copy must be submitted to SEA for redistribution
(under the same restrictions). Someone may also use the sources
in a commercial product, but in this case, a source license fee
is charged and the resulting program may NOT be a general purpose
file-archiving utility.
A short time later PKXARC appeared on the scene, followed
quickly by PKARC. Katz hadn't followed the rules, but then, ARC
wasn't making anybody rich as shareware, and Katz wasn't address-
ing the much more lucrative commercial market SEA had developed
for ARC, so SEA overlooked it. Then, last year, an ad for PKARC
and PKXARC appeared in "PC Tech Journal" on the page facing SEA's
ad for ARC. Katz' ad priced the product a dollar and a half less
than ARC, and even went so far as to make comparisons to "the
other archive utility."
SEA then sent a "cease and desist" letter to PKWare, propos-
ing the following agreement: PKWare would withdraw all commer-
cial advertising and cease attempts at commercial distribution,
and SEA would forgive past violations and grant PKWare an unlim-
ited cost-free license to market its derivative products as
shareware with a non-commercial restriction. Katz refused.
Artikeln är längre men där är ett bra läge att kapa. Så för att summera;
Phil Katz och hans företag PKWARE utgick från sourcekoden till företaget
SEAs ARC komprimerare. Han skapade en kompatibel komprimerare som var
snabbare och bättre och några år senare så stämde SEA PKWARE. De på SEAs
sida pekade på att Katz hade brutit mot copyrightlagarna och de på Katz sida
menade att det här var ett sätt för SEA att bli av med en konkurrent som
gjorde en betydligt bättre produkt. Många blev uppretade över stämningen och
opinionen svängde åt support till Katz håll. (Se också fidonews 605 för mer
information). Det hela ledde också till strävande efter en mer öppen
standard.
FidoNews 5-40, 3 Oct 1988
All we want is a PUBLIC DOMAIN archiving system which is immune to
this nonsense!
En annan kommentar från samma nummer (som var helt tillägnad ämnet SEA vs
PKWARE)
In this controversy, neither side is 100% right, but I hope the
BBS world goes with Phil Katz' new Shareware for several reasons.
PKxARC is much, much faster than ARC. It also compresses some
files more than ARC
[...]
I also understand that Phil Katz was ordered to change the file
extension from .ARC, which he has done. After the first of the
year, however, he can no longer distribute the program. He has
to come up with another file compression algorithm. He expects
to come up with a new algorithm that's superior to ARC. I'm
looking forward to seeing his new program!
Just as ARC has almost completely replaced older schemes like SQZ
and LBR, I hope that Phil's new scheme will completely replace
ARC and become the new BBS standard.
Och ett tag senare hade han ett alternativ till ARC..
FidoNews 6-11, 13 Mar 1989
Just now PKWARE released the shareware program PKZIP. It pro-
duces .ZIP files (called zipfiles) that are not compatible with
ARC. However, it runs as fast as PKARC or even faster, and it
squeezes harder than PKARC. Also, with extra compression turned
on (by typing the options -ea4 -eb4), it squeezes even harder
than NoGate PAK and runs about as fast compressing, and much
faster expanding!
I benchmarked all these programs by compressing version 1.21 of
my ADA-TUTR (Ada Tutor) program, which contains 33 files totaling
more than 700K.
[...]
PROGRAM: COMPRESSED SIZE: COMPRESSION TIME: EXPANSION TIME:
Normal Compression:
PKARC v. 3.5 319577 62 secs. 63 secs.
PKZIP v. 0.9 297045 61 secs. 66 secs.
Maximum Compression:
NoGate PAK v. 1.0 287228 155 secs. 154 secs.
PKZIP -ea4 -eb4 257399 160 secs. 57 secs.
ARC vann slaget i domstolen men förlorade kriget. Förutom att Katz tagit
fram en överlägsen produkt så hade ARC dragit öronen åt sig och nästa
version var inte längre shareware utan en traditionell kommersiell produkt.
Det betydde att spridningen hämmades och att man stötte sig ännu mer med de
som använde deras produkt. ARCs sourcekod hade släppts och folk portade den
till andra plattformar. Men i och med de nya versionerna som införde
komprimering som inte var bakåtkompabtibel förlorade man stödet hos icke-DOS
plattformarna. Det var ytterligare en spik i kistan.
FidoNews 7-34, 20 Aug 1990
Then if you're not running DOS or OS/2 you will not be able to
access the contents of that archive, since there is no source
available for any dearchiver and SEA is only releasing DOS and
OS/2 software. That's intolerable.
1988 var ARC de facto standarden men den tappade mark hela tiden. Klippet
ovan är en del av Fidonews Editor in Chiefs förklaring om varför han gick
över till LHArc som fortsätter..
LHArc is Freely Available with source in both Intel assembler and
ANSI C. No marketing pitch, no fancy sales talk, always up to
date on all platforms. And it's about as good as anything else
(and MUCH better than ARC 5.12!). It's slower than most, but the
bottom line is connect time. Smaller archives cost less to move.
Fri source är tyvärr något som Fidonet inte varit bra på. Hade programmerare
sluppit uppfinna hjulet hela tiden hade vi idag haft ett bättre läge. Det
finns gott om program där skaparen försvunnit och det hela fallit i glömska.
Hursomhelst, resten är historia. ZIP är dominerande än idag på
Windowsplattformen. Det finns bättre komprimerare men idag är det inte lika
viktigt att pressa filstorlek så mycket som möjligt. Tradition och
kompabilitet över plattformar och programvaror väger tungt.
Slutligen vill jag citera Thom Henderson, skapare av ARC, mannen bakom SEA
(och SEAdog, utformningen av det tidiga Fidonews samt medskyldig till
Alternet, IFNA och P4, men allt det där är (minst) en annan historia..)
när han way back in the days yttrade följande:
FidoNews 2-22, 15 Jul 85
I've been writing and selling software for a long time now.
I do this stuff for a living. I, too, have the dream of
coming up with a winner, something that everyone will want
and use.
---
* Origin: -bonka- [08-83 34 96] (2:201/145)
|