17 comments

[ 2.3 ms ] story [ 55.8 ms ] thread
interested that in 1984 the started microcomputer support - that was the same year i started doing that at what is now middlesex university uk - must have been something in the air, though i soon moved up to writing for the dec-10 and vaxen and later the ibm 4381s. best job i've ever had.
I wonder why this timeline has been upvoted to the front page with no comments.

> 1994 In September a Digital Alpha 2100 server with four processors and 1 gigabyte of memory (known as sable) was installed to replace black as the University's general-purpose Unix server.

> 1996 In June a second Digital Alpha server was installed (known as ermine). This was bought because the load on sable was much greater than had been envisaged.

I remember those. They ran OSF/1. At the time i arrived, sable was for staff and graduate students, and ermine was for undergraduates. In the few years i was there, demand for shell accounts declined to much that they retired ermine, and moved the undergraduates over to sable.

(i might have the two machines the wrong way round)

Around this time, OUCS also operated a modem pool, so users could dial in from their homes over the PSTN. The UK national phone network is mostly run by BT, but there could also be other operators who sort of federated with BT. In Oxford, there was NTL, and the OUCS modem pool was on NTL. I had a BT phone, and OUCS being on NTL made no difference. But i knew some guys in a nerd house who had NTL phone service, and it turned out NTL offered free local calls to other NTL phones. So they just had a modem connection to OUCS running 24/7, connected to a machine on their house network configured as a router. That was the first time i came across always-on internet at home; it seemed scandalously radical.

(i might have misremembered the name of the phone company)

It was ntl (styled as ntl:). Like your nerd house friends I had ntl in my house on the Cowley Road, and they offered a dual-line service so we had one line connected to the modem pool full-time (my SuSE Linux desktop ran an RFC1918 network for the other computers in the house).

Sadly around the time this history comes to an end OUCS (now IT Services) abandoned much of its work with bespoke open source development to become the Microsoft money funnel it is now. I say "sadly" not just out of a preference for free software: the way in which this was done deeply affected the mental health of some of the staff involved.

> Sadly around the time this history comes to an end OUCS (now IT Services) abandoned much of its work with bespoke open source development to become the Microsoft money funnel it is now.

Yeah, a lot of UK universities seem to have gone this route (the number who pay huge sums of money to use the atrocious Blackboard software is scandalous). I think they got burnt by their old-school proprietary software being outcompeted by modern software practices in the private sector. But what they really ought to do is create a non-profit collective that is shared between all (or tens of - you wouldn't need that many to make it a huge saving) universities in the UK to write the software they need for them.

I have friends who were in IT at UCL, and yes, it sounds horrifying. All the people who actually knew how to do things for users are gone, and there are hundreds of managers who can't and won't do anything instead.

I'm not sure it was a rational decision based on experiences with proprietary software. It sounds like good old grift to me.

It's not that I disagree with you exactly, but as someone who teaches at a UK university (much less prestigious than Oxford or UCL)... I used to complain a lot about how much money we were sending to Microsoft and Blackboard etc, but then someone pointed to me that the main non-outsourced piece of software we use (the main student records/teaching info system) is borderline unusuable. Horribly slow, lots of 90s errors like commas in course names breaking things, doesn't work on a tablet, etc etc. At least the Microsoft stuff works.
I hadn't used a shell prior to 1993, looking back I'm amazed at how much I learned in my first few weeks using black. Fond memories of using elm for email, tinyfugue to connect to muds and writing my first web pages. I can't remember if it was black or sable that a small group of us crashed a few times with a deliberate fork() bomb. Not our proudest moment looking back!
The 1906A at Oxford in the early 1970s was one of the first computers I used, I was 16 or 17 years old and at school in Oxford. I'm sure this was partly responsible for my subsequent computing career.
In the early 70s the Oxford Physics Dept ran a Decsystem 10 as I recall that isn't mentioned in the history. Unfortunately the only language available through the timesharing system was Basic.
1998: "On August 4th the University's JANET connection was upgraded to a 34 Mbps ATM link."

- or, in modern terms, a domestic ADSL line.

Crazy to think how many 1990s users could be supported by such a line.
Not to be picky or pedantic (ha!) but that should be a SDSL line to be comparable to (synchronous) ATM. That 34Mbps ADSL (A for 'Asynchronous') line probably does no more than 4Mbps on the uplink.
Doesn't the A in ADSL stand for asymmetric?
Yes, it does, got them mixed up. Still, it is the same idea: same rates up and down versus different rates.
Yeah, but you could also just assume some 100-120 down with 30-40 up. Still same "ballpark" for me, and yet better than the upload of my official 50/10 that is actually 60/12 or so.
It is odd that as late as April 1994 the info service was migrated to gopher instead of http. By that time, most of the colleges had their own websites up and running. Students themselves were rapidly authoring web content (my contribution on my college server was a choose your own adventure from my room, past my dusty empty pigeonhole, through town to my common room bar). It was already clear that the ease of contribution meant http was going to dominate going forward. I guess the gopher migration wasn't a quick project.