Is Anything Ever Really New?

[Griffin NoName]

I am suffering.

I used to enjoy reading New Scientist but a trend I noticed a couple of years ago seems to be increasing.

Is it my mind, or is it New Scientist?

Examples:

1. Article entitled Chip Revolution poses problems for programmers

This beefs on about how hard it would be to programme anything over four-core processors and describes the problems being solved by new "inventions" such as TM or X10 which (paraphrase the crucial bit) will only lock those pieces of memory (ie. not more memory than needed) at the time required by each task (ie. not for longer than needed) and giving child processses priority to avoid deadlocks.

I was doing this kind of stuff 15 years ago for "tasks" being run by several people at once and I am sure I wasn't the only one (I'm not that clever). I don't see any difference in concept between people running conflicting tasks and several processors running conflicting tasks. The problem is they all need the information at the same time and the programmer has to consider and resolve that. Old hat.

To me, this article is really saying, programming became a lazy sport and now it may get harder for a bit again, so only clever clogs will survive in the short term.

2. Article : Future Recall

This makes the stunning revelation that one uses the memory bit of the brain to think about the past and the future. Paraphrasing, the brain shows no essential difference doing either.

Am I stupid, or is this so obvious as to be drivel? How could one ever imagine the future without using stuff the brain acquired since birth? ie memories. And isn't it a fair bet that the mechanisms would appear similar when using stored information to imagine stuff (ok I am assuming that remembering is effectively imagining stuff, but to me that has always felt like what I do: "retrieve" info, "run" info).

So why take a whole article and beef on as if it's all incredible. Why not stick to what seems obvious has been proven by MRI?

3. The same type of critique is supplied by my son when space stuff is written about. ie. it's nothing to write home about. For example, that whole argument about Beagle not having enough juice - I vaguely remember something sensational about 60 watt light bulbs - was totally inaccurate in the way it was reported.

Re. almost all articles on computers, I spend my whole time reading going Bah! we knew that 30, 40, 50 .... years ago.

Can I assume that extrapolating, everything in this magazine provokes this reaction from anyone who knows anything substantial about what they are writing about?

Has this got worse over the last 10 years of this magazines history?

Or is my mind just not letting me enjoy reading it as much as I used to?

[beagle]

I might be missing something but I think you're absolutely right on 1.  Designing multiport memory might be a pain for hardware engineers, but for programmers once you get above programming for two threads the issues don't change much IMHO.
Maybe it's an oblique reference to Amdahl's Law being more obvious when the environment has more processors available than the algorithm can use?

Working in a mixed Windows/Unix environment my experience is that Windows programmers seem much more comfortable with multi-processor/multi-threading than UNIX ones. More than once I've been in the position of explaining to someone that, yes the algorithm is very clever, but unless they work out how to do more in parallel they're still going to have to explain to the customer why it's only occupying 25% CPU while they're waiting; i.e. all of only one of four processors.

I never really used to like New Scientist. The  articles seemed too brief and I preferred the in-depth style of Scientific American. If I got all the way through one of those articles I felt properly briefed on a subject.

[Griffin NoName]

I don't see any divide between Unix/Windows. My teeth were cut on Unix in terms of multi-tasking while Windows was still pretty basic. It may be that the Unix programmers you talk to don't actually know how to thrash it as an operating system. This would make sense if those programmers were essentially using Unix to perform clever scientific stuff but not clued up on operating systems?

NB. I failed to make it clear I wasn't writing about you being no better than a 60 watt light bulb. I was of course refering to the failed Beagle project. I apologise if this has caused any distress

[beagle]

I think it's that the Unix programmers I mix with are used to one big processor whereas the Windows ones have been brought up on a twin Xeon/Windows Server environment where the task manager processor cpu display makes it embarassingly obvious if you're single threaded. Your mileage may of course vary...

Ship/dog/spacecraft/light bulb, you name it. My versatility is as boundless as my ego.

[The Meromorph]

I personally think a lot of this is because almost all PC development, GUI development, Windows, DOS, Linux, AIX etcetera started, and continues to have strong roots in a counter-culture to the 'mainframe' environment.
There is a consequent and persistent assumption that all 'mainframe' techniques, standards, philosophies, and even languages primary purpose is to stifle people's creativity; and there is a consequent rejection of both further developing ('building on') existing solutions (particularly if they were developed in the mainframe environment), and of learning from the experience of others.
I constantly see PC guys reiterating, and failing to solve, the mistakes we made in the sixties and seventies, and failing to see that the rejection of standards is, in plain fact, crippling the creativity they think they are trying to protect.

[Sibling Zono (anon1mat0)]

Mmm.. the memory problem with multicore/processors is AFAIK one of performance, the optimization loading and using the cache is the main trick (not duplicating info in the cache of different processors and the interconnection to address it if is not in the local one). It is first a hardware problem but with proper programming optimization it can make all the difference, which is critical if you are running a taxing task.

I haven't read New Scientist, I am subscribed to Scientific American and Discover although I am not renovating the latter because apparently it's having some political interference in it's articles lately. 

[Griffin NoName]

 

I think it's that the Unix programmers I mix with are used to one big processor whereas the Windows ones have been brought up on a twin Xeon/Windows Server environment where the task manager processor cpu display makes it embarassingly obvious if you're single threaded. Your mileage may of course vary...

Yep. Especially if the Unix proggers you mix with never look at what their big processor is actually doing.... the point about Unix v. Windows WAS you had to know how to find those things out for youself, not just know there was a handy tool to flip a window up... and when you did delve into it, you had to know how to do the stuff that gave you the control rather than the processor....

All of which takes me back to my original point

I personally think a lot of this is because almost all PC development, GUI development, Windows, DOS, Linux, AIX etcetera started, and continues to have strong roots in a counter-culture to the 'mainframe' environment
<snip>
......... and of learning from the experience of others.
I constantly see PC guys reiterating, and failing to solve, the mistakes we made in the sixties and seventies, and failing to see that the rejection of standards is, in plain fact, crippling the creativity they think they are trying to protect.

I agree. Sweet memories of ICL George and trying to optimise cluster terminals......

Nowadays the word cluster conjures other images.

EDIT - cross posting...

...... not duplicating info in the cache of different processors and the interconnection to address it if is not in the local one

Yes, but even this is not a NEW idea. We had approaches to the same problems with buffering and for example, the early days of sort theory, extended memory, virtual memory, relative addressing etc etc etc

These issues are real, need improved solutions, etc, but they are not new which is what irritates me with the NS journalism.

Maybe I should change to Scientific American?

[Bluenose]

Yes, its interesting how people often try to deny the lessons learned in years gone by.  I once had to rewrite a major subsystem on a mainframe system (yes, it was written in COBOL).  the original program did a lot of database work which was all very well, but it was slow and for other technical reasons had reached the limit of the amount of data it could handle.  I was given the job of re-engineering it to enable it to cope with an immediate doubling of the data volume and hopefully not imposing a similar (albeit bigger) limit on the new process.

I had a big think about what we were trying to do and realised that the problem was that the original process was continuing to access the DB.  I wrote an extract program and then proceeded down a stream of 5 further programs using flat file processing and instead of making the smarts in my programs, I used an arbitrary key and the heavy duty sort utilities that IBM had provided as part of the operating system, only requiring a line or two of JCL to use.

The end result was a system that processed twice the amount of data each night but in less than half the run time of the original program.  It was an old fashioned approach but in this case entirely appropriate and completely open ended regarding the amount of data that could be processed.  The new ideas are good, but you need to understand where we came from and why certain things were done the way they were.  As has been said before if we do not learn the lessons of history we are doomed to repeat the mistakes.  As applicable in IT as anywhere.

I like the new graphic environments, I like using a GUI to perform admin functions that I don't do very often - if for no other reason than then I don't have to worry about making a simple syntax errors that brings the house down.  However, it does come in handy at times understanding how it all works under the cover.  Even in my old COBOL days, those of us who understood what the computer actually did with the code we were writing were able to get better results for less effort.

No, I am not sure that there ever really is anything new under the sun!

[Griffin NoName]

Waxing nostalgic, anyone remember the effort of getting magnetic tapes to the right position for the next data trawl while parallel processing something else that didn't need that data yet?

Or optimising programs by watching the lights on the mainframe signalling how each assembler code function was being exceuted?

And having rotas timetabled to come in and hose down the machines with cold water according to when they would overheat.....?

And programming line printers to make music?

[beagle]

...Or travelling across Surrey as a 6th former in 1977 to use the Varian computer at Ewell Tech.

Repeat
Load boot sequence on front panel from switches.
Load O.S. from paper tape.
Load program from paper tape.
Crash computer
Until thoroughly_pissed_off && pubs_open

[Griffin NoName]

Arrr Lad, ye were lucky to have paper tape. In my day all we had were them books ye looked numbers up in....

Me favourite were werking fer the NHS. It were 2 weeks in the pub every run. No wunder its in truble.

[Sibling Qwertyuiopasd]

on #2... of course you're imagining the future with the same thing as your past. I mean, if you're remembering something, you're imagining it. You're certainly not experiencing it (usually).

that's why you sometimes remember things differently, or incorrectly. you're imaging in it based on what you know, and people forget things. the only difference between imagining the past and imagining the future is that someone (and somewhere in you) KNOWS what happened in the past. to think about the future you do have to use your 'imagination' and creativity.

I wouldn't know anything about computer programming. or much.

~Qwerty

[The Meromorph]

Bluenose I remember doing a very similar thing on my first job in America. Cut the nightly batch run from 7 hours to 45 minutes. Impressed the hell out of them the first week I was there... 

[Griffin NoName]

Looks like if we all teamed together we could make a bomb sorting out the huge terrible impossible problem described in NS if we could persuade Them to trust people with a few grey hairs

[Bluenose]

Yes, and I've worked out what gives you grey hair too, it's haircuts - everytime I get my hair cut there seems to be more grey hairs, so that proves it!

[Griffin NoName]

And I found out what cures it.

When my hair came back after chemotherapy all the grey had disappeared

It's been 4 years now and still no new grey hairs. But then my body still isn't back to normal functioning either <sigh>

[Bluenose]

I trust you willunderstand my reluctance to use that cure.  I think I'd rather just continue to go grey!

[Griffin NoName]

What you need is a Manproduct !

[Bluenose]

LOL!

I think this might be a new enterprise by Bustlin Brian...

[Griffin NoName]

Bustlin Brian must be a busy man.... seems I can cure my chronic fatigue syndrome along the same lines.

Unbelievable luck 5000 years ago

Ayurvedic medicine is based on ancient natural healing laws that were founded by geniuses...

...You see Ayurvedic medicine was discovered by holy men who acquired this knowledge from God while in the deepest states of meditation.

This knowledge was then recorded into scriptures (which exist today).



And this knowledge was then tested and perfected over 5006 years on millions of human beings making it the most perfect healing system in existence

Note to self: remember to tell Bustlin Brian which day to roll onto 5007 million years of testing

[beagle]

...You see Ayurvedic medicine was discovered by holy men who acquired this knowledge from God while in the deepest states of meditation.

I suppose it would have been impertinent of them to ask why He didn't just fix the design, rather than issue a list of workrounds.

[Bluenose]

Perhaps He works for Westland helicopters.

Back in the 70s the RAN purchased a number of Westland Seaking ASW helicopters.  They went for the Westland version rather than the original Sikorsky one because the Westland version had more powerful Rolls Royce engines and could fly on one engine if necessary.  (Although it had a reduced payload and could carry only 2 torpedoes instead of 4 on the Sikorsky one - go figure.)

Anyway, a few years into their service these helos started to develop cracks in one of the ribs.  Westland was duly consulted and in due course provided a patch to remedy the problem.  By and by, a crack appeared in the next rib, back to Westland, another pathc was supplied.  And so, right through a series of ribs in the aft fuselage of these helicopters.

We later found out that Sikorsky had had the excact same problem way back when they had first produced the type and had modified the design to incorporate all the necessary mods so that new aircraft did not experience the problem.  Westland had bought the original design, but none of the in-service patches that Sikorsky had developed.   

Sibling Bluenose

[Aggie]

Actually (mom is into alternative health, and has read up on many things I sneer at), I place more faith in Ayurveda than most other alternative health systems, at a "gut" level (my gut tells me what to eat to feel good, and this largely jives with Ayurveda).