The application platform’s in play. But where’s it going? This application platform thing looks like it has legs. A few weeks ago we asked the Reg audience at large if they ‘got’ the idea of the application platform, and, practicalities aside, most said ‘we do’. Around a quarter said they had already adopted one. Digging a bit further, we see that of those that are ‘somewhere’ with …
Are application platforms really any different to application servers? Is this just rebranding by Microsoft and its band of merry followers
In any case why am I required to login and provide a whole load of personal details just to listen to the broadcast? This all sounds like a load of M$ shite to me.
The opposite concept would be "one application, one system", which is pretty much a non-starter for larger companies, though it used to be PC dogma (Pournelle's law).
I was once involved with a cross-platform product numbering application, jump-started after a merger. For speed, the original data was set up on a database, the decision which ERP and which groupware to use came later. As different parts sought an interface, we observed that they were connecting to existing copies: the replicate of the database, the ERP copy and the groupware copy, which made them de facto platforms. Smaller customers sometimes made their own copy in a corner of their application.
It is rare to watch platforms grow from seed. but pruning to fit is pretty common.
No remeniscence icon, I'll make do with a smiley - it was a fun project.
Perhaps, a link in this article to the excellent Register article on the emerging Buzzword Application Platform would help.
Which excellent Register article? I don't know that's why I want a link.
Meanwhile I will be at Walmart as I need to buy some walls.
"The opposite concept would be "one application, one system", which is pretty much a non-starter for larger companies, though it used to be PC dogma (Pournelle's law)."
No, "one application, one system" was never PC dogma. Most people didn't know how to make a PC multitask, but it was far from impossible. For example, device drivers were loaded into memory and worked alongside command.com. Early viruses used the same mechanism. Basically, if you could print from Word Perfect or VisiCalc, you were multitasking. I'd go into details on the TSR mechanism, but you're probably already yawning.
The actual quote was "one user, one CPU", in Jerry's Chaos Manor column in Byte Magazine in (probably) 1984ish ... I sold my complete USAian collection to an idiot on Ebay, or I'd look up the exact issue (the shipping costs alone were astronomical ... I'm not sure who was the biggest twit, me for keeping 'em for so long, or the dude who bought nearly a ton of obsolete, unrecycleable clay-coated paper and had it shipped from Palo Alto to Philly ...).
Jerry later updated it to "one user, multiple CPUs" ... He was a very early adopter of multitasking and networking of home computers. Seems to me that he also had something of a "law" that said something about checking the cables first ... But one of my early mentors was constantly harping about "check the wire first!" ... maybe the memories have merged in the mists of time.
You have a fine memory, no need to keep the papers. And you are right, 'one user one CPU' is more exact, says http://en.wikipedia.org/wiki/Jerry_Pournelle.
That said, you are describing fast switching rather than multiuser multitasking, and a system - a nebulous idea in itself - today would surely bundle in the printing and viruses(cookies, codecs) as well as processing and I/O. Data and hardware sharing are the important drivers for platforms, as opposed to each his own spreadsheet or system.
'Dogma' and 'laws' make things sound over-certain, too - my pointing was away from the direction of small machines doing one thing, that doesn't fit well with larger businesses - I had modern database servers in mind as I wrote.
"you are describing fast switching rather than multiuser multitasking"
Yes. But you are the one who brought up PC dogma & Pournelle. I was assuming (I know, assUme and all that ...) that you were talking about IBM compatible/clone PCs in the early/mid '80s. Seemed a reasonable assumption, at the time of writing.
"my pointing was away from the direction of small machines doing one thing, that doesn't fit well with larger businesses - I had modern database servers in mind as I wrote."
Whole 'nuther kettle of fish. Multiple CPUs for a single application (and/or a couple satellite applications), and many local users[1]. I get it. I cut my eyeteeth on mainframes, and later Tymeshare and misc. DEC gear, leading to TOPS-10 and UNIX ... Context and timing is everything[2] ...
[1] "local users" here meaning "this company, perhaps geographically well distributed".
[2] Or so I remember my assembler mentor telling me :-)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
