Home Messages Index
[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index

Re: [News] Another Big Blow to Microsoft in the UK (Schools and Free Software)

Homer wrote:
Verily I say unto thee, that High Plains Thumper spake thusly:

It may be old school, but one does not need a special
software development application to program, if they learn
old school techniques.  With proper technique, it does not
matter what code is used, assembly, C, FORTRAN, ALGOL,
BASIC, COBOL, FORTH, PL1, RPG, and etc.  Coding is
technician work.

As I've said before, for most of my first year at Uni, I
didn't even /touch/ a computer, except for the end of week SML
assignment, which took all of five minutes to copy from my
jotter, and test before printing. I spent most of my lab time
goofing off (i.e. hacking Kerberos auth tickets to make it
seem like my popup Zephyr massages on other workstations were
coming from the department head).

I've always found that trying to actually /learn/ by sitting
in front of a computer is just a distraction, even back in the
days before WWW. I can't imagine that /children/ learn any
faster or better in front of a PC than by listening to a
teacher and doing written assignments. Certainly all my
anecdotal evidence points to that being the case.

Learning is about understanding /principles/, not application.

Seems that schools are turning out script kiddies

I doubt that any of them even know what a script is. Their
definition of "programming" probably involves big shiny
buttons and a rodent.

Sounds like you had a good software programming foundation to build the rest of your career on.

Computer is nothing more than a power tool. If one doesn't understand principle, it serves them no more usefulness than a desktop paperweight.

I remember one senior engineer exclaiming "spaghetti code" with a FORTRAN-VI program with embedded assembly in parts. I took on that "spaghetti code" and studied it, found out it was highly structured. Contracted designer/writer was a genius, but the company did not want to pay him an additional $10K for the documentation. Had one comment for every 50 or so lines of code. Dumped it into WordStar 3.0 on an HP-125 CP/M-80 computer with dual full height floppies after downloading through an RS-232 link from the mainframe.

Formatted the code, adding comments as I discovered nested loops, and-if-then-else's, conditionals, etc. Formed a highly commented data dictionary, etc., when variables and scalars were discovered. (It is challenging to figure out what 6 character names were, example, LTAPEA meant 10 inch reel-to-reel 1600 or 6250 BPI tape was initialised and armed for recording data; LTAPEF meant tape was at EOF marker.) It had even renumerated types ala FORTRAN-IV PARAMETER data statements. One would have never guessed by the column 7 programming style. Compiled it and compared image CRC's, to ensure no changes were introduced by my formatting and commenting.

After several weeks of careful study, was able to ascertain all changes required to the software to make sound changes for upgrades to the channel setups for load servo control and data acquisition to accomodate the newer hardware for aircraft structural testing.

We had about 20,000 lines of uncommented FORTRAN-IV and IBM 360J Assembly code to deal with, added another 5,000.

It would not have been possible if one did not understand principles of structured software design and wasn't willing to learn assembly. Again, coding is a technician's issue.

programming technicians skilled in higher level languages
only.

If one can bear to describe crap like Visual Basic as a "high"
level language.

These are fine for some tasks.

RADs are fine for non-programmers, at companies too cheap to
hire actual software engineers.

When it comes to operating system level programming, one
needs tighter code.

Someone should tell the bozos responsible for the Vista
scheduler ;)

When tight code is executed on a fast computer, then it is
just that much faster.  When operated on a slower one,
execution speed is still acceptable.

Although I rarely attended any, I still remember the Demo
Scene parties in the 80's (still going strong today), where
coders produced the most amazing demos in just 4k of RAM. Pity
the Vole seems to hire the afore -mentioned Visual Basic
"receptionists" instead of ex-sceners, otherwise they might
not produce so much Bloatware.

This is one of the reasons why Linux is much more efficient
at tasks than Windows.  At 2 MB, Beryl is a good example of
efficient 3D desktop work.

It does highlight one of the more pronounced technical
differences between the Linux and Windows development
methodologies, ostensibly at least.

I was able to speed up data acquisition channel scanning from 50,000 to 130,000 aggregate channels per second on this 2 MIP minicomputer with 2 MB RAM, by careful use of embedded assembly to replace FORTRAN library calls in the scan loop and breaking out the tape write routine as a separate task trap triggered by queued intertask communications.

Accorded, I doubt but only a few Windows developers know about assembly language (or how to optimally use low level programming languages like "C". One can create *bloat* with any programming language.)

--
HPT

[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index