Bootstrap ancestry
in [UK Mac]
|
From: Rowland McDonnell on 20 Jun 2010 20:02 Richard Tobin <richard(a)cogsci.ed.ac.uk> wrote: > Rowland McDonnell <real-address-in-sig(a)flur.bltigibbet.invalid> wrote: > > >> I just > >> mean the transitive closure of the programs used to produce it. > > >I can see nothing that permits any meaning to be attached to the > >intersection of the concepts of transitive closure with that of computer > >software. > > Consider the binary relation R on the set of computer programs where > A R B iff B was used to produce A. Let R' be the transitive closure > of R. By the "ancestors" of a program P I mean the set of programs Q > such that P R' Q. I had to look up `transitive closure' when it was first used in this thread. And I still don't get it. Wikip says this: "In mathematics, a binary relation R over a set X is transitive if whenever an element a is related to an element b, and b is in turn related to an element c, then a is also related to c." Okay, but what's this binary relation R all about? <shrug> I don't understand that. I understand elements of a set relating to each other, but I don't fully get what `binary relation R' means. .... but it *is* talking about ordered pairs of members of a set, and I don't see how that relates to the matter in hand. btw, I don't reckon that your definition of `ancestors of a program' is a valid one. Surely it's like saying that any food I eat is my ancestor? Rowland. -- Remove the animal for email address: rowland.mcdonnell(a)dog.physics.org Sorry - the spam got to me http://www.mag-uk.org http://www.bmf.co.uk UK biker? Join MAG and the BMF and stop the Eurocrats banning biking
From: Rowland McDonnell on 20 Jun 2010 20:07 Ian McCall <ian(a)eruvia.org> wrote: > richard(a)cogsci.ed.ac.uk (Richard Tobin) said: > > > Woody <usenet(a)alienrat.co.uk> wrote: > > > >>> Lancaster University at the time I was there ran an experimental Dynix > >>> machine that used 16 386 chips (I forget the megahertz) - this was > >>> considered advanced stuff. > > > >> Wow, 16,386 is a lot :) > > > > Probably a typo for 16,384. > > Nope - 16 physical 386 chips, on a Sequent Symmetry. I believe I've > remembered now that they were running at 16Mhz. cent1.lancs.ac.uk, > although of course then it was uk.ac.lancs.cent1 - the JANET way of > addressing still makes more sense from a routing point of view to me, > but I accept that battle is long gone and also from a memory point of > view the internet way of doing things might be better (as it mimics how > addresses work in the real world - most specific bit of info first). I don't see why they didn't define the addressing so as to be used by the machinery in the fashion most convenient for the machinery, and ditto for us. So store the address one way, and display it the other. And then someone wants to use a plain text editor for preparing emails... Rowland. -- Remove the animal for email address: rowland.mcdonnell(a)dog.physics.org Sorry - the spam got to me http://www.mag-uk.org http://www.bmf.co.uk UK biker? Join MAG and the BMF and stop the Eurocrats banning biking
From: Rowland McDonnell on 20 Jun 2010 20:17 Adrian Tuddenham <adrian(a)poppyrecords.invalid.invalid> wrote: > Justin C <justin.1006(a)purestblue.com> wrote: > > > Adrian Tuddenham wrote: > > > Justin C <justin.1006(a)purestblue.com> wrote: > > > > > >> In article <883cceFgicU1(a)mid.individual.net>, Chris Ridd wrote: > > >> > What sort of blind alleys would we avoid? > > >> > > >> Copper telephone lines to each house - we'd put fibre optic in from the > > >> outset! > > > > > > ...and when the mains fails during an emergency situation...? > > > > Oh, you really are a glass-half-empty kinda guy, aren't you?! > > This is not something to be dismissed with "It'll never happen". Power > cuts aren't all that frequent, but sometimes they occur as the result of > something catastrophic - that is not the time to discover that the > 'phone doesn't work when the mains is off. There is a legal requirement > for one telephone per line to work independently of the mains and there > is a good reason for that rule. And I don't see what's so bad about copper wires for data anyway, shortage of copper aside. It's easy to use 'em for data tranmission using very simple tech - you have to have highly advanced manufacturing processes even to make decent optical fibres, and it's hard to see how you could develop that without having a good telecoms system in place first. And that'd have to be `electrical conductors' unless you want to go for free-space communication using photons (pick yer own frequency). > [We can throw the fire extinguishers away, we didn't use any of them > last year.] I remember a staff meeting I attended at a 6th form college (one I was teaching at). There was much moaning and groaning when my boss - erm, well: `sod, she's started off on one again' as someone muttered. Well, maybe, but: she /was/ safety officer, and the builders the college had got it *were* totally blocking three fire exits with building materials, and fires in educational establishments *do* happen, and people *DO* get burnt horribly to death if the fire escapes aren't all clear... Some people are very very stupid about such issues. Oh yeah, and we did actually have a real fire that year. Even after *THAT*, they didn't get it! ARGH!!!!!!! Rowland. -- Remove the animal for email address: rowland.mcdonnell(a)dog.physics.org Sorry - the spam got to me http://www.mag-uk.org http://www.bmf.co.uk UK biker? Join MAG and the BMF and stop the Eurocrats banning biking
From: Rowland McDonnell on 20 Jun 2010 22:33 Justin C <justin.1006(a)purestblue.com> wrote: > In article <883cceFgicU1(a)mid.individual.net>, Chris Ridd wrote: > > > What sort of blind alleys would we avoid? > > x86? > > I think that knowing, in advance, that while at first 32bit was > liberating, it became a mill-stone, and backward compatibility issues > sure hindered progress. That potential problem was spotted in advance and so was no problem at all for those who had planned for dealing with it. 32 bit is not any sort of mill-stone if you have a properly designed architecture - 32 bit did nothing to hold Macs back at all. It was only what we called the Wintel world got held back by that, surely? And then only because MS insisted on backwards compatibility in a way intended to stifle progress and competition. If MS had not made that mistake, but had instead done it the Apple way - throwing away the old code base entirely, after a transition period involving emulation - you wouldn't be saying that 32 bit was a mill-stone. (I'd say `albatross', anyway - who's met a mill-stone outside of a museum or heritage site in this country?) However, none of that is relevant: if all technological artifacts were removed now, it'd be centuries if ever before humanity got back to building ICs, after a period of such total social catastrophe and near-extinction that hardly any ideas from now could possible survive. So: applying hindsight would be impossible. Remember that the spec for getting rid of all technological artifacts also gets rid of books and libraries... > 64bit good, 128bit better, 256bit? 4 bits best, don't forget. No reason to think that 256 bit is going to be any good, not this side of 2050AD. No reason to think that 128 bits is better than 64 bits with the current state of the art, either. What do I mean, `4 bits best'? 4 bit microcontrollers are very cheap and do the job very nicely in many cases - hence, 4 bits best. The question is `For what job?' when you specify `best', am I right? > Would we try > and find a way to avoid this kind of thing completely? What, by designing a hardware-independent method of coding? Can't see that - if you want to make efficient use of your hardware and you haven't got very very clever software development tools, you need to pay attention to the details of the hardware. > Massive parallelism? To get more bits add another cpu? Not a new idea, surely? Bitslicing was invented years ago. So was massive parallelism. > I think that the architecture could be very different. Lots of different architectures were tried out, but the x86 brand won due to historical issues not closely related to technical issues. To have got a wider range of architectures out here in the marketplace, all you'd've needed to do is to have kept the monopoly-hunters out of it. Basically, the fact that all the interesting different ways of doing the jobs died out in the personal computing world is because of the Wintel take-over. > This is the sort > of question that it would be great to ask of those people who have > worked 8bit 16bit 32bit and 64bit - if you could start again, how would > you do it? The way it was done - all the interesting ideas that could be come up with were come up with. Problem is, hardly any of them survived. They failed to cut it in the marketplace. How come? What a lot of people miss is that the computing hardware's not the important bit: what's important is what you can do with it, which means what's important is *the use to which the users can put the machinery to*. And *that* depends on them having software - /that they can use/. So some hardware's better than others - so? If what you have lets you get your job done, and you've got it, and it works, why change? That's heart of it, I think - what drives people to buy computer hardware is the promise of what they can actually get done with it, not whether or not it's any good under the bonnet. Doesn't matter if computer X is twice the speed if it'd take new software you'd have to buy and then learn - not when you've got computer Y and the required software and you can use it all and you can do the job *NOW*. Billg spotted that a *VERY* long time ago, and his firm wiped out the competition almost completely, laid waste to a marvellous diverse ecosystem, leaving us with rats and cockroaches rummaging through the rubble and weeds, all through applying an unimaginative but clever business lawyer's brain to the task, rather than the brain of a massively creative and optimistic `let's make the cake bigger' techno-utopianist which is what most of the rest of 'em are/were. Rowland. -- Remove the animal for email address: rowland.mcdonnell(a)dog.physics.org Sorry - the spam got to me http://www.mag-uk.org http://www.bmf.co.uk UK biker? Join MAG and the BMF and stop the Eurocrats banning biking
From: Richard Tobin on 21 Jun 2010 04:30
In article <1jkev8o.12wu3bymdwku0N%real-address-in-sig(a)flur.bltigibbet.invalid>, Rowland McDonnell <real-address-in-sig(a)flur.bltigibbet.invalid> wrote: >I had to look up `transitive closure' when it was first used in this >thread. > >And I still don't get it. Wikip says this: > >"In mathematics, a binary relation R over a set X is transitive if >whenever an element a is related to an element b, and b is in turn >related to an element c, then a is also related to c." Example: "ancestor" is the transitive closure of "parent". Given (ordered) pairs of people A and B, some are in the relation "B is the parent of A". "Parent" isn't a transitive relation, because B being the parent of A and C being the parent of B doesn't imply that C is the parent of A. The relation "ancestor" is what you get if you keep applying "parent". It's called the closure because you go on until you don't get any new elements - the resulting set is closed. It's transitive in that if B is an ancestor of A and C is an ancestor of B, then C is an ancestor of A. >btw, I don't reckon that your definition of `ancestors of a program' is >a valid one. Surely it's like saying that any food I eat is my >ancestor? Yes, it's just like that. The food you eat isn't your ancestor, but it seems like a reasonable metaphor. -- Richard |