Re: Effect of SPAM on number of active contributors

I know this started as a complaint about spam content in the Google
version of this Forum.
But I see a need to answer several opinions about the merits of the
F77 Fortran sub-set today.

I respect the various arguments given for defending facilities that
other persons consider important in a Fortran compiler. But I suspect
most work in other fields from myself.

As for the "six spaces" needed per line when writing fixed-form
Fortran, you just type a statement or format label number, or hit the
editor's macro key that puts in the six spaces; and you backspace if
you need a continuation digit. In free form you still need to type
spaces if you want readablity. Or else you can use Microsofts
Development System editor which anticipates you when writing fixed-
form.

As for the supposed benefits of IMPLICIT NONE, I prefer to have my
variable names as immediately identifiable for the type they are under
the default system (and I only use the name LOLD for the single
logical variable I have to use in file operations). And all my
character variables start with C.

Now, the CASE process takes far more tests than a computed GO TO,
especially when you are processing key strokes (four kinds of function
keys? Four kinds of Pad keys? ALT-number? Tab, backtab, backspace,
three kinds of return? Three kinds of arrow keys? Four kinds of
alpanumeric?). And throw in a mouse being used at the same time.
No, I don't have anything good to say for the CASE "facility". And all
this may be happening while typing into a form on screen.

Look, I HAVE written quite a few F90 programs to get native-Windows
"look-and-feel" (to molify some clients) where I needed to call API's.

I do a LOT of text parsing work, and a lot of form data-gathering
work.

Almost all my development programming in Fortran (mainly F77, and
since 1983) has been towards writing suites of table-driven programs,
of general applicability, mostly to process files of data from
experiments, or from commercial surveys (from TV-watching recording
hardware, or the familiar street, Mall, or telephone interviews, or e-
mailed or web-site survey forms). Or even for accounting, data-mining
and auditing purposes. And for anything where a form is on the screen.

Table-driven programming is a very powerful technique. Even you CPU
uses it. The CPU in a computer takes the machine instruction it finds
at the next instruction address, as an operation number and possible
references to result and holding registers and any new "immediate"
entering values. The operation number is essentially a pointer to a
table of steps needed to perform the desired operation on the defined
data, and what to do with the result.

I'll skip all the complications, but several individual table-driven
text parsers are needed in survey processing, to be able to have one
general suite that will process any data, and produce statistical
reports in the original user language, as question texts, column
headings, individual row texts, with appropriate names for totals and
statistical measures and confidence levels and so on.

The main point is that data is accumulated as two dimensional event-
measure arrays, (formed from a table-driven prosess from the input
data) which are then searched with table-driven instructions (to meet
logical conditions) that were again themselves compiled by the text-
processing compiler, and 2 and 3-dimensional matrices are generated
from these selections, as sums or counts,

But never do two or more matrices need to be combined in mathematical
operations.

So: matrix manipulation as such is not needed from the Fortran
compiler.

Then the report writing program (table driven) processes these
matrices into rows of columned text (descriptions and counts or
measures or quantities), usually as a WORD RTF document.

Each of the parser-compiler outputs tables, used by other programs to
execute the indications sub-processes. Maybe 90% of any of my 30 to 60-
program suites are text parsing programs.
.

Relevant Pages

  • Re: F77 vs F90+
    ... | Fortran, you just type a statement or format label number, or hit the ... | Development System editor which anticipates you when writing fixed- ... without the need to touch any keys. ... | mailed or web-site survey forms). ...
    (comp.lang.fortran)
  • Re: Bit Concatenation
    ... use a current compiler, it's generally a F95 compiler, But that really is ... I try to write "standard" Fortran. ... you mention below have nothing to do with F90 versus F77. ... Only about writing new interfaces for nearly two hundred modules. ...
    (comp.lang.fortran)
  • Re: Fortran 77 parser
    ... the difficult involved with one person writing a Fortran 77 ... Actually I have now completed the compiler and it was fiddly but really not ... yielding a kind of AST for each line. ... If anyone is interested in a Fortran 77 to anything compiler then I could ...
    (comp.lang.fortran)
  • Re: Fortran 77 parser
    ... Actually I have now completed the compiler and it was fiddly but really ... Writing a full parser of Fortran 77 ... which supports all syntax is extremely demanding. ...
    (comp.lang.fortran)
  • Re: Fortran 77 parser
    ... Actually I have now completed the compiler and it was fiddly but really not ... Writing a full parser of Fortran 77 ... which converts a "typical" Fortran code or even better a Fortran code ...
    (comp.lang.fortran)