gfortran vs. g95 
Author Message
 gfortran vs. g95

Hi

        I have looked at the websites of both these projects :

        http://www.*-*-*.com/
        http://www.*-*-*.com/

        and g95 just seems to be far more complete (I have compiled some of my
substantial code with it). Why is it then that many people keep
recommending gfortran ? I ask this question not to find fault, but to know
if I have missed something about gfortran ? g95 now has optimization,
according to some reports to be within 10 % of IFC, so that cannot be the
reason.

        Secondly, one of these was spun off the other. Finally, when we have two
open source compilers available (and g95 seems to be closer to that than
gfortran is), what would be the intended selling points of one over the
other ? (As a user, I would definitely not work with both compilers, though
I can understand that another user might choose a different compiler for
some reason.)

Thanks.



Sun, 01 Apr 2007 02:49:20 GMT  
 gfortran vs. g95

Quote:

> Incidentally, if the accusations concerning source code obfuscation are
> true, then an additional problem with G95 is that G95 cannot be legally
> distributed or copied. In some jusrisdictions this probably makes even
> using G95 illegal.

I note that Andy's site still asks people for FSF copyright assignment
forms, although he handles them in a very non-standard way. Just what is
the FSF position on g95? I didn't think it was considered part of their
domain.

--
Scott Robert Ladd
site: http://www.coyotegulch.com
blog: http://chaoticcoyote.blogspot.com



Sun, 01 Apr 2007 21:01:39 GMT  
 gfortran vs. g95
<snip> g95 now has optimization, according to some reports to be within 10 %
of IFC<snip>

It would be interesting to see actual benchmark comparisons. Could you point
towards the reports you mention if they are online somewhere?

I wonder if g95 is on its way into the Polyhedron compiler comparison...

Kind Regards,
Stig Kildeg?rd



Sun, 01 Apr 2007 19:32:25 GMT  
 gfortran vs. g95

Quote:


>> My understanding is that the reason is not technical but personal. Read
>> the information at
>> http://gfortran.org/index.php/Gfortran/TheOtherGccBasedF95Compiler

>Amazing. I guess personality clashes can cause problems even in online
>collaborations.

>> Incidentally, if the accusations concerning source code obfuscation are
>> true, then an additional problem with G95 is that G95 cannot be legally
>> distributed or copied. In some jusrisdictions this probably makes even
>> using G95 illegal.

>> (Standard disclaimer - I am not a lawyer but I do have a law degree).

>> David

>Well, if it was true then, it does not seem to be true now (depending upon
>what is meant by obfuscation above). There are instructions on the g95
>website on how to compile it from source. If the source were to be somehow
>modified from its actual content, that process could not possibly work.

I have not looked at the G95 source and cannot comment on it.

In general, source code is said to be "obfuscated" when a program compiled
from it produces the same results as the original, but the source has been
made artificially difficult to read. You could obfuscate source code by removing
all comments and indentation, renaming all variables to i01,i02,i03, etc.
(Some programmers do this by habit :).) Sometimes source code is deliberately
sold in an obfuscated form. For example, the Math a la Carte library is sold
in "mangled" format -- see http://mathalacarte.com/license.html . It is still
standard F77 code that the user can compile on any platform, but it is deliberately
made unreadable. It is intended to serve as portable object code. Human readable
code for individual algorithms can also be purchased, at a higher price.

----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =---



Sun, 01 Apr 2007 21:21:17 GMT  
 gfortran vs. g95
On 13 Oct 2004 08:21:17 -0500

Quote:



> >> My understanding is that the reason is not technical but personal.
> >Read> the information at
> >> http://gfortran.org/index.php/Gfortran/TheOtherGccBasedF95Compiler

> >Amazing. I guess personality clashes can cause problems even in
> >online collaborations.

> >> Incidentally, if the accusations concerning source code obfuscation
> >are> true, then an additional problem with G95 is that G95 cannot be
> >legally> distributed or copied. In some jusrisdictions this probably
> >makes even> using G95 illegal.

> >> (Standard disclaimer - I am not a lawyer but I do have a law
> >degree).>
> >> David

> >Well, if it was true then, it does not seem to be true now (depending
> >upon what is meant by obfuscation above). There are instructions on
> >the g95 website on how to compile it from source. If the source were
> >to be somehow modified from its actual content, that process could
> >not possibly work.

> I have not looked at the G95 source and cannot comment on it.

> In general, source code is said to be "obfuscated" when a program
> compiled from it produces the same results as the original, but the
> source has been made artificially difficult to read. You could
> obfuscate source code by removing all comments and indentation,
> renaming all variables to i01,i02,i03, etc.(Some programmers do this
> by habit :).) Sometimes source code is deliberately sold in an
> obfuscated form. For example, the Math a la Carte library is sold in
> "mangled" format -- see http://mathalacarte.com/license.html . It is
> still standard F77 code that the user can compile on any platform, but
> it is deliberately made unreadable. It is intended to serve as
> portable object code. Human readable code for individual algorithms
> can also be purchased, at a higher price.

That's a pretty good explaination of obfuscation. In this case, in
addition to being rude, if the source code is obfuscated, it does not
qualify as distributing source for the purposes of the GPL since the GPL
defines source as: "the preferred form of the work for making
modifications to it." Clearly this refers to the unobfuscated version of
the source.

David



Sun, 01 Apr 2007 21:36:21 GMT  
 gfortran vs. g95

Quote:
> I saw this report here :

> http://www.g95.org/

> Scroll down to entries for August 30 :

> "Douglas Cox reported a major speedup in the code that g95 generates,
> within
> about 10% of ifort on -O3. All the -fxxx optimizations (f is for funky)
> should also work as well."

> I guess you could contact Mr. Cox above for the benchmarks if he has them.

Thank you for the info.


Sun, 01 Apr 2007 20:41:00 GMT  
 gfortran vs. g95

Quote:

> Hi

>         I have looked at the websites of both these projects :

>         http://www.g95.org/
>         http://gcc.gnu.org/fortran/

>         and g95 just seems to be far more complete (I have compiled some of my
> substantial code with it). Why is it then that many people keep
> recommending gfortran ? I ask this question not to find fault, but to know
> if I have missed something about gfortran ? g95 now has optimization,
> according to some reports to be within 10 % of IFC, so that cannot be the
> reason.

>         Secondly, one of these was spun off the other. Finally, when we have two
> open source compilers available (and g95 seems to be closer to that than
> gfortran is), what would be the intended selling points of one over the
> other ? (As a user, I would definitely not work with both compilers, though
> I can understand that another user might choose a different compiler for
> some reason.)

gfortran.org links to some earlier messages indicating how this came about.

Since most people (or me at least :-) still hope for reconciliation between
both projects, we tend not to talk much about this subject publically, and
I'll let it be with that reference to stuff that's already public. Unless
someone wants to start a flamewar of course :-)

If you have questions about gfortran you're welcome to look at gfortran.org,

answered there.

Regards,
- Tobias Schlter
(g95 contributor when it still accepted contributions, now gfortran contributor)



Sun, 01 Apr 2007 05:08:38 GMT  
 gfortran vs. g95

Quote:



>>> My understanding is that the reason is not technical but personal. Read
>>> the information at
>>> http://gfortran.org/index.php/Gfortran/TheOtherGccBasedF95Compiler

>>Amazing. I guess personality clashes can cause problems even in online
>>collaborations.

>>> Incidentally, if the accusations concerning source code obfuscation are
>>> true, then an additional problem with G95 is that G95 cannot be legally
>>> distributed or copied. In some jusrisdictions this probably makes even
>>> using G95 illegal.

>>> (Standard disclaimer - I am not a lawyer but I do have a law degree).

>>> David

>>Well, if it was true then, it does not seem to be true now (depending upon
>>what is meant by obfuscation above). There are instructions on the g95
>>website on how to compile it from source. If the source were to be somehow
>>modified from its actual content, that process could not possibly work.

> I have not looked at the G95 source and cannot comment on it.

> In general, source code is said to be "obfuscated" when a program compiled
> from it produces the same results as the original, but the source has been
> made artificially difficult to read. You could obfuscate source code by
> removing all comments and indentation, renaming all variables to
> i01,i02,i03, etc. (Some programmers do this by habit :).) Sometimes source
> code is deliberately sold in an obfuscated form. For example, the Math a
> la Carte library is sold in "mangled" format -- see
> http://mathalacarte.com/license.html . It is still standard F77 code that
> the user can compile on any platform, but it is deliberately made
> unreadable. It is intended to serve as portable object code. Human
> readable code for individual algorithms can also be purchased, at a higher
> price.

Thanks for the explanation.

I know a few colleagues who "obfuscate" by habit ! I remember a piece of
code which I was incorporating into my own and found that my friend had
used variables like axm=Planck's constant, riom=effective mass, etc. And
the person abhorred all comments, and did no indentation ! It was a
nightmare for me to go through the code.

So, isn't the definition of obfuscation pretty vague ? Some programmers
(including myself many many years ago) just have bad (IMO) naming
conventions. Even now, I sometimes use variables like temp1, temp2, etc.
when I just need the information to be passed between subprograms through a
module (quick and dirty way). Its not my usual practice, but sometimes, no
logical name occurs to mind.

Unless one can find two sets of sources - one on which the person in
question worked (with sane names, etc.) and another in which a search and
replace strategy seemed to have been employed, I think one should not
confuse bad (IMO) programming practices with obfuscation.

Someone just might have a different idea of "the preferred form of the work
for making modifications to it". Does the GPL define obfuscation for its
purposes a little more precisely ?



Mon, 02 Apr 2007 03:50:24 GMT  
 gfortran vs. g95
On Wed, 13 Oct 2004 15:50:24 -0400

Quote:




> >>> My understanding is that the reason is not technical but personal.
> >Read>> the information at
> >>> http://www.*-*-*.com/

> >>Amazing. I guess personality clashes can cause problems even in
> >online>collaborations.

> >>> Incidentally, if the accusations concerning source code
> >obfuscation are>> true, then an additional problem with G95 is that
> >G95 cannot be legally>> distributed or copied. In some jusrisdictions
> >this probably makes even>> using G95 illegal.

> >>> (Standard disclaimer - I am not a lawyer but I do have a law
> >degree).>>
> >>> David

> >>Well, if it was true then, it does not seem to be true now
> >(depending upon>what is meant by obfuscation above). There are
> >instructions on the g95>website on how to compile it from source. If
> >the source were to be somehow>modified from its actual content, that
> >process could not possibly work.

> > I have not looked at the G95 source and cannot comment on it.

> > In general, source code is said to be "obfuscated" when a program
> > compiled from it produces the same results as the original, but the
> > source has been made artificially difficult to read. You could
> > obfuscate source code by removing all comments and indentation,
> > renaming all variables to i01,i02,i03, etc. (Some programmers do
> > this by habit :).) Sometimes source code is deliberately sold in an
> > obfuscated form. For example, the Math a la Carte library is sold in
> > "mangled" format -- see http://www.*-*-*.com/
> > still standard F77 code that the user can compile on any platform,
> > but it is deliberately made unreadable. It is intended to serve as
> > portable object code. Human readable code for individual algorithms
> > can also be purchased, at a higher price.

> Thanks for the explanation.

> I know a few colleagues who "obfuscate" by habit ! I remember a piece
> of code which I was incorporating into my own and found that my friend
> had used variables like axm=Planck's constant, riom=effective mass,
> etc. And the person abhorred all comments, and did no indentation ! It
> was a nightmare for me to go through the code.

> So, isn't the definition of obfuscation pretty vague ? Some
> programmers(including myself many many years ago) just have bad (IMO)
> naming conventions. Even now, I sometimes use variables like temp1,
> temp2, etc. when I just need the information to be passed between
> subprograms through a module (quick and dirty way). Its not my usual
> practice, but sometimes, no logical name occurs to mind.

> Unless one can find two sets of sources - one on which the person in
> question worked (with sane names, etc.) and another in which a search
> and replace strategy seemed to have been employed, I think one should
> not confuse bad (IMO) programming practices with obfuscation.

> Someone just might have a different idea of "the preferred form of the
> work for making modifications to it". Does the GPL define obfuscation
> for its purposes a little more precisely ?

The GPL doesn't define obfuscation. The gist is that you have to give
out the version of the sources that you usually edit.  You're not
allowed to{*filter*}around with them to make them difficult to work with
before you give them out. That means that you're right, you need to be
able to show that that's going on. I am not a g95 or gfortran
contributor so see Tobias Schlueter's comments on what the gfortran
people think is going on here.

Also note the smiley i the previous poster's post - the GPL does not
outlaw crappy programming.

David



Mon, 02 Apr 2007 16:20:31 GMT  
 gfortran vs. g95

Quote:

> I saw this report here :

> http://www.g95.org/

> Scroll down to entries for August 30 :

> "Douglas Cox reported a major speedup in the code that g95 generates, within
> about 10% of ifort on -O3. All the -fxxx optimizations (f is for funky)
> should also work as well."

Well, I haven't studied g95's code base since about a year and a half,
but in general, the -fxxx optimizations are implemented in code that
isn't touched by either g95 or gfortran (it's either in the GIMPLE
optimization passes or even after that).

Perhaps it helps if I explain the various parts of g95/gfortran:

1. Parse Fortran 95 in free and fixed format - mostly written by Andy
2. Perform name resolution on the resulting internal representation -
    mostly written by Andy.
3. Convert the internal representation into the top-level intermediate
    (i.e., language independent) representation of GCC (called GIMPLE)
    and implement scalarisation of array operations - written by Paul
    Brook and Steven Bosscher.
4. Perform optimizations on the GIMPLE-representation, such as loop
    optimizations and vectorization - a huge number of people from
    Red Hat, SuSE, IBM Israel, and various others [here are -fxxx
    optimizations].
5. Convert the optimized GIMPLE to the low-level internal representation
    known as RTL (Register Transfer Language), a quasi assembly language
    for a processor with an indefinite number of registers - Jason
    Merrill of Red Hat and various others.
6. Perform optimizations on the RTL representation [here are some more
    -fxxx optimizations] - old, various contributors from Richard
    Stallman in the eighties to several people from different
    organisations nowadays.
7. Convert the RTL representation to the final assembler output for the
    specific processor - this is very old and still has original code
    in it from the early stages of GCC (i.e., before 1990) - Richard
    Stallman, Richard Kenner, Jim Wilson and others.

I hope this explains that everyone in this gfortran/g95 effort *has* to
stand on the shoulders of giants in order to get anything useful done.

--

Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
A maintainer of GNU Fortran 95: http://gcc.gnu.org/fortran/



Mon, 02 Apr 2007 05:01:47 GMT  
 gfortran vs. g95

Quote:


>>Incidentally, if the accusations concerning source code obfuscation are
>>true, then an additional problem with G95 is that G95 cannot be legally
>>distributed or copied. In some jusrisdictions this probably makes even
>>using G95 illegal.

Can't comment on that (not being an attorney at law).

Quote:
> I note that Andy's site still asks people for FSF copyright assignment
> forms, although he handles them in a very non-standard way. Just what is
> the FSF position on g95? I didn't think it was considered part of their
> domain.

We decided to stop looking at g95's source code, because the way the
updates are described makes it impossible to be sure what code is
contributed by whom.

Those who think that's not important are welcome to read up on about one
and a half year of discusing the SCO vs IBM case on Groklaw
(www.groklaw.net) explaining why it is.

--

Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
A maintainer of GNU Fortran 95: http://gcc.gnu.org/fortran/



Mon, 02 Apr 2007 04:36:27 GMT  
 gfortran vs. g95

Quote:


>>Incidentally, if the accusations concerning source code obfuscation are
>>true, then an additional problem with G95 is that G95 cannot be legally
>>distributed or copied. In some jusrisdictions this probably makes even
>>using G95 illegal.

>>(Standard disclaimer - I am not a lawyer but I do have a law degree).

>>David

> Well, if it was true then, it does not seem to be true now (depending upon
> what is meant by obfuscation above). There are instructions on the g95
> website on how to compile it from source. If the source were to be somehow
> modified from its actual content, that process could not possibly work.

Since I wrote that sentence on gfortran.org, I guess I'll have to explain it,
even though I wanted to stay out of this thread. Since Andy Vaught doesn't
keep a public history of g95's code (which IMO counts as obfuscation by
itself), this might change any second, so it might not be obfuscated any more
when you look.

There are three types of obfuscations in the publically accessible sources:
1. random whitespace added at ends of lines (makes diffing sources difficult,
easily cured with indent)
2. order of functions randomly interchanged (makes diffing difficult, and the
code somewhat harder to read).
3. variable names randomly changed (makes diffing impossible, and the code
much harder to read)

The latter seems in part to be done automatically, i.e. there are variable
names that change systematically (I haven't looked at the code in a while, but
I think 'e' -> 'a' always, or something like that), and there are others where
human work is obvious (as seen in one place: 'first' -> 'next', 'next' ->
'last'). You can also see that this is not the code Andy is using by comparing
the line numbers from the debug info with the sources, but as before, this can
change any second.

(There are also only very few comments, which is more of a style thing, but it
also makes the code harder to reads)

Since by the GPL it's legally possible to move code from g95 to gcc (of which
gfortran is a part), and since these obfuscations can only be aimed at
preventing this, it should give you an idea how Andy Vaught feels about gcc,
the compiler he's using.

- Tobias Schlter

I will stay out of this thread from now on, promised :-)



Mon, 02 Apr 2007 01:39:02 GMT  
 gfortran vs. g95

Quote:


>> <snip> g95 now has optimization, according to some reports to be within
>> 10 % of IFC<snip>

>> It would be interesting to see actual benchmark comparisons. Could you
>> point towards the reports you mention if they are online somewhere?

>> I wonder if g95 is on its way into the Polyhedron compiler comparison...

>> Kind Regards,
>> Stig Kildeg?rd

> I saw this report here :

> http://www.g95.org/

> Scroll down to entries for August 30 :

> "Douglas Cox reported a major speedup in the code that g95 generates,
> within about 10% of ifort on -O3. All the -fxxx optimizations (f is for
> funky) should also work as well."

> I guess you could contact Mr. Cox above for the benchmarks if he has them.

I believe the comments quoted above were about the results obtained using
the mdbnch.f program, which on my system produces the (partial) output
shown below.

Doug Cox

Summary of Results for ifort: [-O3]

    MDBNCH: A MOLECULAR DYNAMICS BENCHMARK, VERSION OF DECEMBER 17, 1988
<snip>
    10 TIME STEPS,    2 LIST UPDATES
    0.004999 SEC. CP TIME FOR INITIALIZATION
    0.101985 SEC. CP TIME UPDATING THE LIST        (  0.050993 SEC/UPD. )
    0.408936 SEC. CP TIME CALCULATING FORCES       (  0.040894 SEC/STEP )
    0.247961 SEC. CP TIME FOR TIME INTEGRATION     (  0.024796 SEC/STEP )
    0.021000 SEC. CP TIME FOR OTHER TASKS          (  0.002100 SEC/STEP )
    0.779882 SEC. CP TIME EXCLUDING INITIALIZATION (  0.077988 SEC/STEP )
    0.784881 SEC. TOTAL CP TIME

 *******************************************************************************
 COMPLETE BENCHMARK EXECUTION TIME :    2.883562 CP SECONDS.

Summary of Results for g95 [-O3]

    MDBNCH: A MOLECULAR DYNAMICS BENCHMARK, VERSION OF DECEMBER 17, 1988
<snip>
    10 TIME STEPS,    2 LIST UPDATES
    0.004999 SEC. CP TIME FOR INITIALIZATION
    0.110984 SEC. CP TIME UPDATING THE LIST        (  0.055492 SEC/UPD. )
    0.437932 SEC. CP TIME CALCULATING FORCES       (  0.043793 SEC/STEP )
    0.253961 SEC. CP TIME FOR TIME INTEGRATION     (  0.025396 SEC/STEP )
    0.022997 SEC. CP TIME FOR OTHER TASKS          (  0.002300 SEC/STEP )
    0.825874 SEC. CP TIME EXCLUDING INITIALIZATION (  0.082587 SEC/STEP )
    0.830873 SEC. TOTAL CP TIME

 *******************************************************************************
 COMPLETE BENCHMARK EXECUTION TIME :    3.167519 CP SECONDS.



Mon, 02 Apr 2007 10:16:09 GMT  
 gfortran vs. g95

Quote:
> if the source code is obfuscated
> David

This discussion wont lead to something revealing, I guess (so sorry
for adding to the noise), but at least looking at the latest source
code (starting from www.g95.org follow 'compilation notes' leading to
http://g95.sourceforge.net/g95_source.tgz) might help to form an
opinion about the level of obfuscation present e.g. a sample routine
(select.c) :

/* SELECT CASE statement
   Copyright (C) 2000, 2001, 2002 Free Software Foundation, Inc.
   Contributed by Andy Vaught and Steven Bosscher

This file is part of G95.

G95 is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2, or (at your option)
any later version.

G95 is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with G95; see the file COPYING.  If not, write to
the Free Software Foundation, 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA.  */

/* select.c-- Handle the SELECT CASE statement. */

#include <string.h>
#include <stdlib.h>

#include "g95.h"

static int overlap;

/* g95_match_select()-- Match a SELECT statement */

match g95_match_select(void) {
g95_expr *e;
match r;

  r = g95_match_label();
  if (r == MATCH_ERROR) return r;

  r = g95_match(" select case ( %e )%t", &e);
  if (r != MATCH_YES) return r;

  new_st.type = EXEC_SELECT;
  new_st.expr = e;

  return MATCH_YES;

Quote:
}

/* match_case_eos()-- Match the end of a case statement */

static match match_case_eos(void) {
char name[G95_MAX_SYMBOL_LEN+1], *block_name;
match i;

  if (g95_match_eos() == MATCH_YES)
    return MATCH_YES;

  g95_gobble_whitespace();

  i = g95_match_name(name);
  if (i != MATCH_YES)
    return i;

  block_name = g95_current_block_name();
  if (strcmp(name, block_name) != 0) {
    g95_error("Expected case name of '%s' at %C", block_name);
    return MATCH_ERROR;
  }

  return g95_match_eos();

Quote:
}

/* compare_case()-- Compare two case nodes.  This is used to sort the
 * nodes into order and also gives us the perfect opportunity to check
 * for overlap.  The default case is sorted first in the list,
 * followed by a unbounded upper, followed by the full ranges in
 * order, followed by the unbounded lower.  We only report the first
 * overlap. */

static int compare_case(const void *f, const void *k) {
[....]



Mon, 02 Apr 2007 04:09:25 GMT  
 gfortran vs. g95

Quote:



>>> My understanding is that the reason is not technical but personal. Read
>>> the information at
>>> http://gfortran.org/index.php/Gfortran/TheOtherGccBasedF95Compiler

>>Amazing. I guess personality clashes can cause problems even in online
>>collaborations.

>>> Incidentally, if the accusations concerning source code obfuscation are
>>> true, then an additional problem with G95 is that G95 cannot be legally
>>> distributed or copied. In some jusrisdictions this probably makes even
>>> using G95 illegal.

>>> (Standard disclaimer - I am not a lawyer but I do have a law degree).

>>> David

>>Well, if it was true then, it does not seem to be true now (depending upon
>>what is meant by obfuscation above). There are instructions on the g95
>>website on how to compile it from source. If the source were to be somehow
>>modified from its actual content, that process could not possibly work.

> I have not looked at the G95 source and cannot comment on it.

> In general, source code is said to be "obfuscated" when a program compiled
> from it produces the same results as the original, but the source has been
> made artificially difficult to read. You could obfuscate source code by
> removing all comments and indentation, renaming all variables to
> i01,i02,i03, etc. (Some programmers do this by habit :).) Sometimes source
> code is deliberately sold in an obfuscated form. For example, the Math a
> la Carte library is sold in "mangled" format -- see
> http://mathalacarte.com/license.html . It is still standard F77 code that
> the user can compile on any platform, but it is deliberately made
> unreadable. It is intended to serve as portable object code. Human
> readable code for individual algorithms can also be purchased, at a higher
> price.

Thanks for the explanation.

I know a few colleagues who "obfuscate" by habit ! I remember a piece of
code which I was incorporating into my own and found that my friend had
used variables like axm=Planck's constant, riom=effective mass, etc. And
the person abhorred all comments, and did no indentation ! It was a
nightmare for me to go through the code.

So, isn't the definition of obfuscation pretty vague ? Some programmers
(including myself many many years ago) just have bad (IMO) naming
conventions. Even now, I sometimes use variables like temp1, temp2, etc.
when I just need the information to be passed between subprograms through a
module.

Unless one can find two sets of sources - one on which the person in
question worked (with sane names, etc.) and another in which a search and
replace strategy seemed to have been employed, I think one should not
confuse bad (IMO) programming practices with obfuscation.

Someone just might have a different idea of "the preferred form of the work
for making modifications to it". Does the GPL define obfuscation for its
purposes a little more precisely ?



Mon, 02 Apr 2007 03:48:46 GMT  
 
 [ 15 post ] 

 Relevant Pages 

1. automatic arrays with negative size in g95 and gfortran

2. Common statement does not work with gfortran, but g95 gives the correct answer

3. gfortran or g95

4. difference between g95 and gfortran

5. g95 AND gfortran problem ONLY on PPC OSX Tiger

6. Compiler flags for compiling FEM2DLiB with g95 and/or gfortran

7. problem with very small numbers - g95 versus gfortran

8. gfortran, g95, and dual-core

9. g95 versus gfortran

10. Gfortran 2 years behind G95 and still not ready for prime time

11. Vector instruction support in G95 or Gfortran?

12. recl on g95 and gfortran

 

 
Powered by phpBB® Forum Software