what's a callback?

There's a difference between doing something because it can be done and doing something because it's needed. Microsoft can throw the problem of hardware resources over the wall to the users at no cost to themselves.

This is precisely the thinking that's needed and necessary for /embedded/ computing. Reliability and hardware resources are directly impacted by keeping things as simple as possible.

"John Larkin" schreef in bericht news: snipped-for-privacy@4ax.com...

Well, you write, test, approve and forget ;) Most of the code is pretty harmless and straight forward anyway.

There's a lot of things behind the scenes. Nobody is amazed when he/she hits the print button and a printer dialog pops up. And I have no idea how *your* printer dialog looks like. And neither has a windows application, when it happens. And the printer driver has no idea where your printer is, at lpt1, com2, usb or somewhere on your ethernet network. And of course you expect a gentle warning that it might be turned off, or has run out of paper or ink. And you want to run it from 7 applications at the same time, each using different papersizes, while your secretary wants the same printer to print a sheet of labels that she has put in the manual feed tray...

That is just one example. It is a lot easier to write robust code for a full featured pinball machine.

Steve at fivetrees" schreef in bericht news: snipped-for-privacy@nildram.net...

going

as

Is

SED is a mess, and a jolly one at that, so a bit of cross-contamination won't hurt. I got a bit pissed by CBfalconer plonking Genome, SED's mascotte. The folks at CAE could loosen up a bit, increasing their value at the same time. Speff even dropped CAE in his follow-ups from a few subthreads in this one, never seen him do that on SED. The atmosphere at CAE is of the arrogant, boring wannabees etc.

;)

The silly thing is that the thread has 'run off' into a whole variety of code areas not associated with a callback at all... A 'callback', is exactly what it says, and comes from the useage on the telephone, where you ring somebody, and they say they will 'call back'. There are a number of reasons. The first would be the obvious one (I'll call back, when I have the answer). The advantage here is it removes the need for an 'interrupt' signal to give a fast response, or a polling process to monitor the response of the subroutine. It can also be used for a security application (just as on the phone, where your bank 'calls back', to verify that they are talking to the person expected). This is rare in PC applications, but is quite common in some other OS's, but can be used on the PC, so that the 'callback', is operating at a different security level to the calling code. There is also the 'billing' reason (as in the phone call, where you call somebody, and they say they will 'call back', so that the charges go onto their account). In the computer version, this allows the 'callback', to be using time from the subroutine process, rather than from the main code. Again this could be at a different priority to the calling routine.

Best Wishes

No, that's just code.

Re-entrant code can be run from multiple threads, and are not dependent upon one thread running the code to completion before the other thread can run the same code.

Typically re-entrant code has no side-effects, and no static variables. Everything modified typically is in registers or on the stack, (or perhaps passed as a parameter by the caller).

All ISRs should be re-entrant, though they needn't be, if you control your interrupt sources.

Re-entrant code isn't simply an ISR. When threads are swapped preemptively, any routine that could be interrupted and called by another thread needs to be re-entrant.

I guess that's the same thing I said in the first paragraph.

Rufus

.... snip ...

Here is the complete message when I PL>> Even if all the snow were burnt, ashes would remain.

Interesting that you provide this brilliant reply to the only response to the original that actually was intelligent. PLONK.

Heh - I run Win 3.11 on a 200MHz machine with 8 gig hard disk and 64 meg ram. Most of the time I'm in DOS, where I use my own operating system. It responds instantaneously to any command and rarely crashes except when Windows calls a bad pointer and takes everything down. But a reboot only takes seconds and I'm back in business.

Your system needs a 2GHz machine with gazillion gigabytes of hard disk and more ram than most disk drives had in the 80's. Your operating system and word processor take much longer to load, and you are constantly attacked by viruses, trojans, adware, and popups.

Tell me again how things have improved since the 80's:)

(Although I do have to admit I will be happy when I can find a 2GHz machine that will run all my code and still use my keyboard and mouse. Adobe is just too slow.)

Best,

Mike Monett

john jardine wrote: [...]

Microsoft has peaked and can only lose market share to Linux. Open Source is becoming a viable business model. Things are improving, but remember, Rome wasn't burned down in a day:)

Best,

Mike Monett

How about this analogy?

I often need to shave dimes and pennies off of hardware for production purposes (corresponding, of course to a multiple of that in the selling price), but when I buy a *tool* for *business use*, an extra $10 or $100 is often no big deal if I get something valuable in return.

Spending a certain amount of today's bountiful desktop disk space and processor processor bandwidth to get a tool that's as feature-rich as makes sense to the authors of the program (hopefully) and easy to install and configure (again) is not silly at all. Even on inexpensive products, we sure don't write everything in tightly crafted assembler any more. It's not as simple as possible, but rather as simple as makes sense with the resources at hand.

Best regards, Spehro Pefhany

"Mike Monett" schreef in bericht news: snipped-for-privacy@spam.com...

Windows isn't bloated at all. There's just a lot of things to be dealt with, but that happens so incredibly transparent to the casual observer, that we take it all for granted and assume the code is bloated or something.

This style of thinking would probably kept us all back in 1988, the release date of your editor ;)

Actually, it's a complete operating system that runs on top of DOS, the same as earlier versions of Windows. Here's a partial screen shot:

.. 2F87BA86 Previous Directory IE50REG 2F87BA86 Backup Copy of MSIE REG16x2.DAT REGDATA 2F87BA87 Explorer 5.0 Verisign Certificates 93% slack SYSTEM 2F87BA88 System Files ZIP 2F87BB39 Backup Directory NISTIMEW CFG 310B6F05 31 Configure NIST Time LMOUSE COM 186A1940 34,658 Logitech Mouse WIN COM 28FA792A 44,170 Win 3.1 Original ANIMOUSE DAT 30FB555D 28 AniMouse Config REG DAT 3196691D 55,922 Win 3.1 Original REG16X2 DAT 2CCD8EFA 58,730 Explorer 5.0 Keep a backup copy in IE50REG NISTIME DIF 3193BD23 1,079 Nist Time Log WINSOCK DLL 20489086 30,516 Moved Here From G:SYMPATIC HK DOC 2EC67370 9,757 HotKey ANIMOUSE EXE 1ED12800 378,464 AniMouse CALC EXE 186A1940 43,072 Win 3.1 Accessories CALENDAR EXE 186A1940 59,824 Win 3.1 Accessories

The columns show the file or directory name, extension, date in hex format, filesize, and a comment field that contains information about the file. This information goes with the file when I move or copy it to a different directory, and the information can be changed as needed without breaking links to other programs, such as image links in html files.

The cursor (not shown) hilites a file in yellow. Pressing various function or command keys will erase the file, copy or move it to a different directory, or execute a specific program associated with the file. For example, it loads an editor or file viewer as needed, or starts Windows and loads the appropriate program all with a single keystroke.

I have numerous search programs that can index the entire disk in minutes and locate any file in seconds. I never have to type in a filename or directory, and all the needed commands are under control of the cursor and function keys. I never have to remember all the options needed with various programs, such as PKZIP or PKUNZIP. They are all encoded and attached when the program loads.

This method is the opposite of the Windows procedure, where you load a program first, then search for the desired file. It is several orders of magnitude faster than Windows, and completely avoids the confusion and errors when you cannot locate the desired file or load the wrong one.

IMHO, this is the way Windows should have been designed from the beginning.

Best,

Mike Monett

That might be understandable if you didn't also managed to indirectly insult other contributors, who posted well before the 20th when you posted this:

No netlag excuse here.

Your newsreader is distorting your view of the world, I think. In a properly threaded view of this thread, this message by Mr. Falconer

which has you so upset is about the *6th* in the entire thread (and 4 levels deep) not the 20th. If it's 20th in yours, that means you're using inappropriate tools, and should be very careful using their results as arguments.

Terminally sick newsreaders are even less of one. Just because Web-based "forums" that so many newbies prefer over USENET generally don't manage to get this anywhere nearly right, doesn't mean we should degrade USENET in a similar way.

Oh, how loudly wintel laugh on their way to the bank. What honestly puzzles me though is, how have the computing/programming industries avoided natural selection forces over the past 10-20 years?. The present monopoly situation would like it that way but it seems only a part answer. Where have all the free thinkers gone. The radicals. The companies/individuals who can spot major failings and grab an opportunity?.

For example. Selection pressure works in the electronics industry and these groups. A well applauded answer/solution to some need or discussion on a complex requirement, would be one that involves great elegance, subtlety, vast knowledge and ends up using just a few transistors and a diode. Anyone can enjoy the result, few can appreciate the process of getting there.

Look at the programming groups. A well applauded answer to an obvious question is one that is comprehensible only within higher ranks of the priesthood and consumes at least half a ream of paper. (eg the above callback 'paper'). There seems no individuals capable of or willing to break rank. Why was Acrobat version 6 allowed to see daylight?. Where's the ace programmers with a vision and a bad-arsed attitude?. I get a distinct smell/echo of a self supporting, arts type 'philosophy' philosophy at work.

regards john

I read in sci.electronics.design that john jardine wrote (in ) about 'what's a callback?', on Wed, 22 Dec 2004:

I don't think we know for sure why the software business shows all the signs of a species about to go extinct through spending too many resources on lavish body decoration (read 'features that hardly anybody will ever use'). It may be that corporate purchasing is to blame; it is easier to get authorisation for $10 000 than for $100, because if it costs $10 000, it MUST work, whereas for $1000 it MUST be too cheap. And forget $100 entirely - that must be some sort of evil program that will cause the company to collapse.

Even back in the 80s, lots of people were calling attention to the poor quality of PC programs compared with, particularly, BBC Micro and Acorn programs. No doubt RISC had something to do with that, but there was a pride of achievement in producing compact code. The market for these computers and their software appreciated that, too, because while memory and storage were not hugely costly, funds were often very short and many computers had minimal memory and disc capacity.

The people writing high-class compact code now seem confined to the same freeware and shareware market sectors.

Fred Bloggs wrote in news:41C5F993.9090009 @nospam.com:

No, re-entrant code has a specific definition, that is not it, others have covered it well enough in this thread.

something.

Nobody forces you to use the latest windows software. If DOS3.22 works for you, why upgrade?

For simple middle-of-the-road embedded computing, yes.

release

ram. Most

calls a bad

back in

more ram

processor

trojans,

Well, I couldn't care less how much GHz or Gigabytes my machine has. I pay the $500 and take the box home. It works and I don't run out of space. Booting takes a minute, most applications start in less than 5 seconds except Protel that shuffles for a minute, automatically reloading the last jobs I am working on. On average I reboot once a day, typically when I have been using Protel a lot.

My editors have syntax highlighting, I can cut and paste between applications, I can view my pdf-files, either from the internet or from my own collection, I can print my own documents as pdf's, print the stuff on the printer in the other room, backup my files to the other computer in my small network, design my printed circuit boards and mail them to my manufacturer, use other graphical programs to make some really good looking pictures, write software for windows and a couple of microcontrollers such as 8051, PIC, AVR, SX, program them with my nifty device programmer, debug them in-circuit, have silly talks with my brother using skype, keep my website up-to-date, running my database, acces my bank account, the list is endless.

machine that

too slow.)

Always try to be happy ;)

program

you

Seems that you've got yourself organized well. But under windows I too select files and hit enter. Or click it with the right mouse key, and choose from print, mail-to, compile/make, copy, zip, or whatever is a valid option for that particular type of file under my mouse pointer. Win3.1 never attracted me. Much too much trouble with that. But a lot has improved since then.