Vous êtes ici : John’s Cornelisse COMPUTER2CASTER diary > 11-02-2004: testing and controlling...
The First European Monotype University, Geneva, Swiss, 13-20 june 2004     
  Report 2004

About the caster     
  Here is text about caster description

History of computer to caster     
  MacMono Interface

John Cornelisse’s Computer to caster interface     

John’s Cornelisse COMPUTER2CASTER diary     

Monotype Corporation: an history     
  The Monotype Chronicles by Lawrence W. Wallis

Monotype Special sorts and Unicode     
  How to bring Monotype special sorts in the Unicode system ?

The Monotype system     
  Monotype specimen book
  Monotype units


11-02-2004: testing and controlling...

Friday February 13th, 2004, par John Cornelisse

The last weeks I gave not much entries in this diary. Busy programming. Very exiting sometimes, though only for a tiny few of us.

A lot of the work consist of controlling, controlling all. Because if not tested, it will not work.

The main problem is, to control all and everything.

I read a book some time ago, in it the author claimed: no program can be tested in all circumstances. And in all its properties. Because of that, it is not possible to proof the all time correctness of a program. The same would apply for hardware. Hardware is so complex, that no human mind can comprehend all the inside features and properties.

If this is true, we can be pleased with all programs do for us. Nowadays computer-programs guard our safety in many instances. whether we travel by train, plane of car. It does not matter. All our society rely on computers nowadays.

I had some problems with my programming last days.

I’m building the very hart of the program: the interpreter. And this routine is large. Too large already was the decision of my compiler. All kinds of cryptic messages appeared on my screen. No real cause, because most lines I type in C, they do not contain much syntax errors, or other crimes against C or logic.

The massages compilers give you, when compiling stops... Those do not excel in clearness. At the best they indicate the direction you can try to find an answer. Most times the only thing to do is look in the lines ahead.

The other thing is: do only small steps, and recompile after every 5 or ten lines. If something goes wrong. It will be probably in the last typed lines.

But the growing program needs more and more segmenting. One file is not enough any more, a large include-file stores all routines tested before. All kinds of basic routines like reading the set of a character, or the line-width, only the routines I’m working on will be recompiled each time.

By trial and error I’ve more or less determined the maximum size a routine can have. That’s many pages, but this evening I will find a way to chop it in a few parts. Out of heap-space. I’ve seen that message too often for a long time.



Dans la même rubrique :
16-02-2004: C-compilers
12-02-2004: Mac or PC ?
03-03-2004: extension manual, producing margins
22-01-2004: 24 Volt direct current
30-01-2004: indentify the port
12-01-2004: the day after
17-01-2004: about Monotype code
31-01-2004: reprogramming IC
13-02-2004: can’t cope with file’s above 64k
10-01-2004: the day before

Thursday November 23rd, 2017
Mail us! | Site Map | Home Computer to Caster | Newsgroup | Admin

Association pour le Patrimoine Industriel - Association Lettres et Images - Imprimerie des Arts