Xah Lee's Unixism

Brian Inglis Brian.Inglis at SystematicSW.Invalid
Tue Sep 7 15:57:24 CEST 2004


On Mon, 06 Sep 04 11:23:17 GMT in alt.folklore.computers,
jmfbahciv at aol.com wrote:

>In article <rv1hhc.mtv2.ln at via.reistad.priv.no>,
>   Morten Reistad <firstname at lastname.pr1v.n0> wrote:
>>In article <413af268$0$19706$61fed72c at news.rcn.com>,
>> <jmfbahciv at aol.com> wrote:
>>>In article <20040904.2231.57679snz at dsl.co.uk>,
>>>   bhk at dsl.co.uk (Brian {Hamilton Kelly}) wrote:
>>>>On Thursday, in article
>>>>     <41371e5c$0$19723$61fed72c at news.rcn.com> jmfbahciv at aol.com
>>>>     wrote:
>>>>
>>>>> In article <2mmdj0t6mjgif88en11skbo3n8uiuj46nc at 4ax.com>,
>>>>>    Brian Inglis <Brian.Inglis at SystematicSW.Invalid> wrote:
>>>>> >MS has been borrowing code from Unix to create a real OS: TCP/IP;
>>>>> >NTFS<-ffs; memory mapped files<-mmap.
>>>>> 
>>>>> All right.  Now I'm mystified.  Why did they have to borrow code
>>>>> from Unix?  They already had VMS.  ISTM, VMS had all of the 
>>>>> above.
>>>>
>>>>VMS (originally) most decidedly did NOT have either TCP/IP or NFS.
>>>
>>>I thought VMS did get TCP/IP into it.  I don't know anything about
>>>NFS.
>>
>>VMS was too early, and was made too politically correct.
>>
>>TCP/IP was NOT politically correct until around 1996 or so. 
>>TPTB wanted OSI, GOSIP/Decnet Phase 5 and all that crud, until we
>>Internet people hammered them. 
>>
>>>>Indeed, it took many years before DEC [sorry, by then it was already
>>>>d|i|g|i|t|a|l] had a TCP/IP stack available for VMS --- the dreaded heap
>>>>of quivering jelly created by the Eunice idiots.

I was never aware that DEC offered TCP/IP. 

>>>>Before that, people who needed TCP/IP on a Vax used various third-party
>>>>solutions, such as the implementations from Carnegie-Mellon (CMU) 

The commercialized product from T[he]W[ollongong]G[roup] and TGV
MultiNet seemed to be ubiquitous. 
Both products seemed to operate application level protocols very much
in leaf node store and forward mode, rather than supporting routing or
passthru, on VMS. 

>>>Sigh!  If CMU had it, I would have assumed it got hornshoed into
>>>VMS.
>>
>>Wrong mindset. TCP/IP was never a DEC invention, much less a D I G I T A L 
>>one. 

IIRC there was a majority of PDP-10s running on ARPAnet (1982 and
earlier) and later TCP/IP (1983 on), but most may have been running
Tenex, as BBN was running the network. 

>It didn't have to be a DEC invention.  If it was CMU, we got it
>shoved down our throats and up our asses.  However, I see
>that the dates explain why TCP/IP didn't get into VMS.  

Politics and not timing was why TCP/IP didn't get into VMS:
d|i|g|i|t|a|l backed the European horse that never ran as it fitted
better with their network hardware capabilities and DECnet plans.
It also meant they did not have to deal with those BBN guys that had
developed a competing OS and network. 
They had whole suites of products layered on top of DECnet that were
sold to European governments and contractors. 
They bet that the ISO and governments couldn't be wrong and they
wouldn't lose out, but they did, as did IBM with SNA networks. 

>Apparently the protocol got good after Gordon Bell left DEC.

TCP/IP didn't get better, but the implementations of OSI networking
performed badly and did not interoperate, so TCP/IP swept the
networking competition off the board, and that may have had an
influence on his departure. 

>Since TCP/IP was in the 90s, I couldn't have heard about it
>over the wall (I think I stopped working in 1987).  I could
>swear that cybercurd meant something.

WWW was in the 90s, as was allowing commercial access to and
competition to operate the Internet backbone, so it became a must have
for the previously clueless, like digital and MS. 

>ISTR, the -20 types yakking about it.

BBN and Tenex heritage probably. 

-- 
Thanks. Take care, Brian Inglis 	Calgary, Alberta, Canada

Brian.Inglis at CSi.com 	(Brian[dot]Inglis{at}SystematicSW[dot]ab[dot]ca)
    fake address		use address above to reply



More information about the Python-list mailing list