Panel Discussion on Information Centres

Transcribed from the printed issue; not yet reviewed. Read it in the PDF of the issue, page 46

Q: How do the so called ‘4th Generation Languages’ fit into the Information Centre strategy?

RC: Can you be sure they will work 5 years on? IBM have landed plenty of people in the cart with VSPC — they could just as well do it with FOCUS if they chose.

FP: “APL is the only language anything like a 4GL!” (James Martin); what does 4GL mean anyway, and are such languages capable of wide adoption?

RW: They are Report Writers to deliver tools to a user in a way which does not distract him from the task in hand. They are tending to move to more ‘English-like’ syntax, and products like Focus certainly have a long life ahead of them.

MG: 4GLs are still a DP-oriented product. APL allows users to do their own thing; 4GLs are productivity aids for DP (RW: yet the salesmen aid them at users?!)

TM: If NOMAD, SASS and the rest don’t run in 5 years’ time, a lot of users really are in trouble!. Many companies already have a variety, and find it hard to link them — APL can come in handy as ‘glue’.

DP: This proliferation may be a barrier in itself — maybe a genuine English ‘front-end’ will soon be necessary?

Q: There is still a DP backlog, with many projects not being tackled …. why??

RC: Faster hardware hasn’t helped much; the human resource is still the limiting factor. We should look seriously at APL even for very large DP systems (Gigabytes upwards).

FP: Programming may be quick in APL, but for big projects the programmer’s time is only a smalll fraction; often APL programmers are lousy at project management, which doesn’t help. If the 4GLs don’t deliver the trend will continue — the user demands function, not a set of arbitrary rules to learn. No-one bothers how VisiCalc works — they just see it!

Q: What about data; do we provide duplicate copies in all sorts of different formats?

FP: The user doesn’t care. He needn’t know where the data is, and in general you should convert it when needed; you can’t keep 6 copies of a real-time database up to date.

TM: At the moment there is no satisfactory answer on IBM systems. Maybe DB2 will help ….?

RC: … it probably will, but golly will it mop up the CPU!

RW: If you have spent 10 years building a set of systems the physical structures are likely to be totally wrong for the IC. Copying may be the only short-term solution.

RC: VS APL finds data access and resource control hard — hence the recommendation to put it on a separate machine. Sharp APL can share files sensibly on the main machine.

FP: 4GLs have this problem too — problems of double updating and integrity can still force seperation and copying.

Q: What effect will the Macintosh have? it looks more suitable for a coffee table?!

TM: People want windowing, but no-one quite knows how it will work with large volumes of data.

??: Nothing an IBM 3279 can do remotely resembles the Macintosh.

??: Are touch screens a gimmick — after a month people prefer to type anyway because it’s a lot quicker?

??: The Sunday supplements have a lot answer for — they have boosted expectations for hardware and increased demand, but done nothing to increase resources.

??: Some of us are just as guilty. In the hands of a real expert APL is a superbly powerful modelling tool — much faster than the 4GLs. Hence a sniff of it raises the user’s expectations, duly increasing that DP backlog!