List of software programs written in fortran (for engineers and scientists)
Jason
(or fortran variant) that was (or is being) used for science and
engineering. I was wondering if someone could help me with
my list..? Here is what I found so far:
1. HEC-RAS U.S. Army Corps of Engineers 1991 River Channel Water
Surface Profile
2. MODFLOW USGS 1983 3D Ground-water flow model
3. ANSYS Dr. John Swanson 1970 finite element analysis, structural
analysis
4. HFPF U.S. EPA Hydrology & Water Quality Water Bodies
5. EPANET Lewis Rossman, U.S. EPA 1993 Dynamic Pipe Flow Distribution
6. SMSIM David M. Boore, USGS 1996 Ground Motions from Earthquakes
7. SPICE1 Laurence Nagel, Donald Pederson 1972 Analog electronic
circuit simulator
..
I would also appreciate any links to references to lists like this
that might already exist.
Thank you.
Louis Krupp
There have been thousands of scientific and engineering programs written
in Fortran. Is there a point in having a necessarily arbitrary and
incomplete list?
Louis
rusi_pathan
impossible to list them all. Just typing Fortran as a topic in ISI's
Web of Science gives 7000+ hits
If you need some examples in different areas then here are some:
- Lots of big climate codes are mostly in Fortran. For example MM5,
WRF, RegCM3, CCSM (POP) etc.
- Many large codes in Geosciences like PFLOTRAN (subsurface/reservoir
flow), SPECFEM3D (wave propagation), GeoFEM (Earth Simulator's
multiphysics code) etc.
- I think Sandia's CTH (shock code) is mostly in Fortran and so was
DYNA3D
- Numerous CFD codes including industrial ones such as StarCD
- Mesh generators such as LANL's Lagrit etc. Mesh adaptation codes
such as PARAMESH, Pyramid etc.
- Solvers/libraries like MUMPS, Scalapack, Blas/Lapack, Sparskit are
used under so many things that it is impossible to even make an
educated guess.
- John Burkardt's webpage has a list of many other libs/utilities
http://people.sc.fsu.edu/~burkardt
Btw arent ABAQUS and parts of Fluent in Fortran too?
Richard Maine
> Jason wrote:
> > I'm writing a paper on software programs originally written in Fortran
> > (or fortran variant) that was (or is being) used for science and
> > engineering. I was wondering if someone could help me with
> > my list..?
> There have been thousands of scientific and engineering programs written
> in Fortran. Is there a point in having a necessarily arbitrary and
> incomplete list?
Thousands? Heck, that's probably conservative. Depending how one counts,
I might claim to have written a few hundred myself. Sure my count goes
down a lot if one insists that they be "major" programs in some sense,
but still...
No, I don't see the point in making a list of even the ones I have done.
I'm sure I couldn't do a complete one, in fact. The details of many
smallish programs that I did in my earlier years are quite lost in my
memory..... and it's not as though it would be useful for me to make
such a list anyway.
I'm thinking that an attempt to make a list of scientific/engineering
programs that have been written in Fortran is either pointless or of
negative value. If the OP thinks it is even within the realm of
possibility to make a list of even "significant" programs, then I'd say
he pretty much misses the whole concept. If anyone read such a list and
believed it, then they would be seriously mislead (thus the negative
value).
Note, by the way, that Fortran majorly predates the web. Thus there are
plenty of Fortran programs that were once in use but you won't readily
find data about on the web at all. Sure, there are ones that were major
enough that you'll find citations of them in later material that was on
the web (and there are even plenty of pre-web programs that are still in
use). But there is quite a large number of older programs about which
you are unlikely to find data readily available on the web.
Maybe if the OP had something more narrow in mind, it might make more
sense. But if that is the case, the first step is accurately expressing
what the real subject matter is; after all, he'll need to do that in
writing a paper anyway.
--
Richard Maine | Good judgment comes from experience;
email: last name at domain . net | experience comes from bad judgment.
domain: summertriangle | -- Mark Twain
Jinsong Zhao
Almost all quantum chemistry programs were written in Fortran. QCPE
stored most earlier version of those programs.
Regards,
Jinsong
Luka Djigas
<jason.li...@gmail.com> wrote:
>I'm writing a paper on software programs originally written in Fortran
>(or fortran variant) that was (or is being) used for science and
>engineering. I was wondering if someone could help me with
>my list..? Here is what I found so far:
>
>1. HEC-RAS U.S. Army Corps of Engineers 1991 River Channel Water
>Surface Profile
>2. MODFLOW USGS 1983 3D Ground-water flow model
>3. ANSYS Dr. John Swanson 1970 finite element analysis, structural
>analysis
NASTRAN, http://en.wikipedia.org/wiki/Nastran
to add at least one.
What exactly do you need a list like this for ?
pp, Luka
viper-2
> I'm writing a paper on software programs originally written in Fortran
> (or fortran variant) that was (or is being) used for science and
> engineering. I was wondering if someone could help me with
> my list..?
Jason:
Everyone knows this, but to make the point, the name "FORTRAN" is
derived from The "IBM Mathematical Formula Translating System". In
fact, some say that FORTRAN is short for "FORmula TRANslator".
The language is practically synonymous with numeric and scientific
computing. Accordingly, I would expect that the number of programs
written in FORTRAN outside the field of science and engineering are a
small minority.
It might be better to find a more meaningful topic for a paper.;-)
agt
--
Freedom - no pane, all gaiGN!
Code Art Now
http://codeartnow.com
Email: a...@codeartnow.com
Luka Djigas
<visi...@mail.infochan.com> wrote:
>On Aug 10, 12:05 am, Jason <jason.lillywh...@gmail.com> wrote:
>> I'm writing a paper on software programs originally written in Fortran
>> (or fortran variant) that was (or is being) used for science and
>> engineering. I was wondering if someone could help me with
>> my list..?
>
>Jason:
>
>Everyone knows this, but to make the point, the name "FORTRAN" is
>derived from The "IBM Mathematical Formula Translating System". In
>fact, some say that FORTRAN is short for "FORmula TRANslator".
>
>The language is practically synonymous with numeric and scientific
>computing. Accordingly, I would expect that the number of programs
>written in FORTRAN outside the field of science and engineering are a
>small minority.
>
>It might be better to find a more meaningful topic for a paper.;-)
>
>agt
Nicely put.
If one may suggest, maybe it would be better to differentiate between
fortran *programs* and fortran *applications*. I don't know how
english speakers differentiate between the two, but I usually consider
an application to be a bigger, and usually gui oriented program (That
is, not saying, that a command line program cannot be big).
So if an OP is looking for a list of fortran applications in view of
big software packages, it would be good to specifically mention that.
pp, Luka
Jason
question. I'm trying to compare Fortran to other early languages that
are still in use today. I want to provide some examples to show why
the languages are used and for what purpose. So I am making a list of
examples that demonstrate an idea, not a comprehensive list of every
program. I've read that Cobol is good for business, Lisp for AI, and
Fortran for science/engineering. I want to show examples that
illustrate that (or maybe something different). I will make sure that
my audience understands that this list is a sampling only.
My current list of fortran programs so far is mainly confined to civil
engineering tools (except for Spice1) - I am a civil engineer, and I
am familiar with these. I thought if I ask this question here, I would
get a more well-rounded response that would be more representative.
Jason
> big software packages, it would be good to specifically mention that.
Yes, I mean fortran applications, not code libraries or programs not
used much. I would also consider commandline driven programs as
applications as long as their contribution was significant.
dpb
Virtually all nuclear design codes both for commercial and defense purposes.
_Small_ sampling would include
PDQ/HARMONY -- neutron diffusion/isotope depletion combination
RELAP -- LWR reactor safety (LOCA, etc., ...) analysis
KENO -- Monte Carlo particle transport
ANISN, DOT3,5 -- Discrete ordinate particle transport
QAD-CGGP2 -- point kernel shielding
ORIGEN -- Isotope Generation and Depletion
The Radiation Safety Information Center (RSIC) at ORNL (Oak Ridge
National Laboratory) is the distribution center for NRC-supported codes
for computations in the field that will have abstracts of these and many
more.
EPRI (formerly known as Electric Power Research Institute) has developed
many software packages for a wide range of problems for the electric
utilities from fossil-fired plant thermal performance monitoring (PMW)
to transmission/distribution network modeling.
Don't know particular codes but at one time many of the petroleum
exploration/production companies geophysical modeling of production
fields were Fortran on CDC/Cray as would run into those guys at a couple
of the timeshare facilities way back when...they were much more
closed-mouth about things than anybody else I ever ran into, though,
other than the most general of generalities so information in that area
is probably going to be much harder to find.
--
Pascal
Jason <jason.li...@gmail.com> a écrit :
Crystallography:
shelx: http://shelx.uni-ac.gwdg.de/SHELX/
crystals: http://www.xtl.ox.ac.uk/crystals.html
Platon: http://www.cryst.chem.uu.nl/platon/
superflip: http://superspace.epfl.ch/superflip/
Pascal
e p chandler
> I'm trying to compare Fortran to other early languages that
> are still in use today. I want to provide some examples to show why
> the languages are used and for what purpose.
IMO the answers to this question are partly historical. The short
version is that early programming languages were designed that way.
COBOL has facilities to manipulate character data, database files and
decimal arithmetic. Fortran specializes in numeric computation. Of
course both langues have evolved in time as well.
> My current list of fortran programs so far is mainly confined to civil
> engineering tools
Perhaps it would be more enlightening to discover what was NOT written
in Fortran, and why? [Not that I want to start a Fortran versus C
flame war....]
But in the end, you may, like Russell, write a very long book that
simply explains fundamentals.
-- e
OldSchool
> ....much. I would also consider commandline driven programs as
> applications as long as their contribution was significant....
"command line driven" as opposed to what? Reading a DB? a flat file?
Cards? GUI?
for example Synercom Technology's AutoSteel read "cards" and did
structural steel design..
glen herrmannsfeldt
< I'm sorry for not being more specific and for asking a misleading
< question. I'm trying to compare Fortran to other early languages that
< are still in use today. I want to provide some examples to show why
< the languages are used and for what purpose. So I am making a list of
< examples that demonstrate an idea, not a comprehensive list of every
< program.
If I wanted a short representative list, I would use those
that are in SPEC. SPEC has some Fortran and some C programs,
meant to represent the wide range of programs people generally use.
I believe that some work went into selecting those, and I wouldn't
want to try to repeat that work, if it wasn't needed.
-- glen
none
sample.
Programs that are *now* no longer written in Fortran still show their
Fortran heritage - Matlab being probably the best example.
Programs that are written in programs that are written in Fortran, or
where the Fortran heritage still shows - such as ACSL.
Many of the wastewater programs - STOAT, Plan-It STOAT, CAPDET for sewage
treatment, OTTER for water treatment, SWMM up to Version 5, InfoWorks, for
river/sewer analysis, WATNET for water networks.
Aspen & PRO/II for chemical engineering process design.
As a rough approximation,
Up to early 1980s: if it was a significant engineering program it was in
Fortran. Algol 60, followed by PL/I then Algol 68, were advanced as the
natural succesor to Fortran, but never managed to take over. In the
mid-1970s the generator programs, such as ACSL, were sold as the future -
Fortran underneat for efficiency, but a supposed 'English like' interface
to speed up solving systems of equations/ODEs/PDEs.
early to late 1980s - Pascal began to take over Fortran share.
early 1990s - Pascal programs began to be unmaintained.
early 1990s to mid 1990s - C began to take over Fortran/Pascal share
late 1990s - C programs began to be unmaintained
late 1990s to date - C++ began to take over Fortran share.
Fortran in the period early 1980s on has had Pascal/C/C++ as declared
succesor programs, with each wave nibbling away at Fortran, and the
Fortran share shrinking - but existing Fortran programs doing as well, or
better, for continued usage and maintenance.
We now appear to be seeing C# in the Windows world replace C++ - while the
Fortran side soldiers on.
No one doubts that there are better languages than Fortran for many tasks.
Personally I felt that Algol 68 was a worthy successor, but it didn't take
off - & I was using iut in the late 1980s.
It is not clear what you are planning to do with a long list - perhaps
better would be to indicate what fields are still writing NEW in Fortran,
rather than continuing to use, or building on, old software. But I feel
that Fortran may be significant even in that arean, as the current
'popular' language tends to be displace completely, while Fortran
continues to retain its small share of new-build.
nm...@cam.ac.uk
OldSchool <scott...@macys.com> wrote:
ICES, anyone? :-)
Virtually ALL of the applications up to about 1990 were command-line
driven, in pretty well any language on pretty well any platform.
There were sometimes GUI front-ends, but they were almost always
optional. Most HPC applications are still like that, for good
reasons.
Regards,
Nick Maclaren.
nm...@cam.ac.uk
none <no...@none.net> wrote:
[ A good summary ]
>early to late 1980s - Pascal began to take over Fortran share.
Grrk. Not really. It was mainly kiddies and computer scientists.
Pascal was dire for practical programming, especially for programs
that used arrays seriously.
>No one doubts that there are better languages than Fortran for many tasks.
>Personally I felt that Algol 68 was a worthy successor, but it didn't take
>off - & I was using iut in the late 1980s.
Yes, I agree. I was using it from about 1970 on. However, Fortran
has now lifted several of Algol 68's innovative ideas ....
Regards,
Nick Maclaren.
Ron Shepard
<deeb8c30-f1a1-4f5e...@f10g2000vbf.googlegroups.com>,
e p chandler <ep...@juno.com> wrote:
> Perhaps it would be more enlightening to discover what was NOT written
> in Fortran, and why?
I think the historical answers to this would almost always involve
interfacing to some kind of graphics library or windowing system or
animation library. This is where fortran programming fell short in the
80's, and the time lag of over a decade between f77 and f90 did not
help. In the 80's the main interaction between man and machine shifted
from batch mode card input with line printer output to an interactive
terminal input with display screen output. Even now, if you want to
interface to, say, an openGL library or a VRML display, it is much
easier to find documentation and tutorials in other languages than in
fortran.
$.02 -Ron Shepard
GaryScott
wrote:
> In article
> <deeb8c30-f1a1-4f5e-93f9-8c1427330...@f10g2000vbf.googlegroups.com>,
True if you must go native, but much easier relatively full featured
solutions for GUIs are available, much easier to program than native,
and at least partially cross-platform portable if you restrict
yourself to the portable subsets.
jfh
> In article <pan.2009.08.10.19.01.36.857...@none.net>,
Another reason why Pascal was dire for practical programming: its
standard specified so little that every compiler I encountered had
many extensions that one had to use. Result: whenever a program had to
be run on a different machine large chunks of it had to be rewritten.
John Harper
Richard Maine
> Another reason why Pascal was dire for practical programming: its
> standard specified so little that every compiler I encountered had
> many extensions that one had to use.
Yeah. I recall running into many people who thought that "Pascal"
implied Borland's product. Borland did have a nice compiler. I bought
(quite a respectable number of version of) it myself and played with it
substantially at home. It was never even tokenly a replacement for
Fortran, though; being limitted to PCs pretty much ruled that out. In
those days, a PC did not count as much of a "serious" computer.
I used it wrote a personal finance program for my own use (before you
could buy such things commercially, or maybe it was just that I wanted
the excuse to play with it - I forget). I also wrote a database program
for keeping track of my personal science fiction library. Don't recall
that I ever did anything for work with it. In fact, I don't think I had
a license for it at work.
captain...@gmail.com
[snip]
> Crystallography:
>
> shelx:http://shelx.uni-ac.gwdg.de/SHELX/
>
> crystals:http://www.xtl.ox.ac.uk/crystals.html
>
> Platon:http://www.cryst.chem.uu.nl/platon/
>
> superflip:http://superspace.epfl.ch/superflip/
>
> Pascal
I am kinda like the Crystallography Source Code Museum at
http://sdpd.univ-lemans.fr/museum/
TideMan
Ahhh Turbo Pascal....
I had a fun project using it to develop a program I called OrFin -
Orienteering Finish - for processing finish times at orienteering
races. When a competitor crossed the finish line, the laptop operator
hit F8 and typed in the competitor's number, then the machine recorded
the time off its clock located them in its database, and subtracted
their pre-set start time to give their race time.
We built it to get around the problem that humans are very bad at
doing sums like subtract 10:56:00 from 12:03:47. Competitors know
exactly how long they took (from their stop watch) and officials get
roasted if they get it wrong, which is very easy when you're
processing lots of finishers (and perhaps you're getting ready to
compete yourself and don't really care).
Pascal
"Captain...@gmail.com" <captain...@gmail.com> a écrit :
Thanks, I was searching for the url and could not find it :)
Pascal
J. F. Cornwall
We use Fortran (77 with extensions) for our water database. "We", in
this case, being the US Geological Survey. It's the National Water
Information System, NWIS, and it's a distributed (? not sure if
technically the right term) database system running on about 50 or so
Sun servers around the US. Each server runs an independent copy and it
all feeds a national Web portal/database server that feeds a subset of
the data to John Q. Public.
The Fortran is in the individual DB servers, as a suite of programs that
are used to ingest, process, export, and manage water-related data. We
use embedded SQL and an Ingres RDBMS. There is a fair amount of C mixed
in as well, but probably 75% is Fortran. We have command-line, GUI, and
old ANSI-escape-sequence "menu" programs in the mix. Code ranges from
brand-new programs, to routines that are older than my college-graduate
son...
Jim
Michael Prager
I've not seen mention of biology so far. My ASPIC program (and
related items) for assessing exploited wild fish stocks has been
used internationally for over 15 years and is still in use.
http://www.sefsc.noaa.gov/mprager/aspic.html
--
Mike Prager, NOAA, Beaufort, NC
Address spam-trapped; remove color to reply.
* Opinions expressed are personal and not represented otherwise.
* Any use of tradenames does not constitute a NOAA endorsement.
robin
news:bbcc40f6-b383-4845...@y4g2000prf.googlegroups.com...
> I'm writing a paper on software programs originally written in Fortran
> (or fortran variant) that was (or is being) used for science and
> engineering. I was wondering if someone could help me with
>
> 1. HEC-RAS U.S. Army Corps of Engineers 1991 River Channel Water
> Surface Profile
> 2. MODFLOW USGS 1983 3D Ground-water flow model
> 3. ANSYS Dr. John Swanson 1970 finite element analysis, structural
> analysis
> 5. EPANET Lewis Rossman, U.S. EPA 1993 Dynamic Pipe Flow Distribution
> 6. SMSIM David M. Boore, USGS 1996 Ground Motions from Earthquakes
> 7. SPICE1 Laurence Nagel, Donald Pederson 1972 Analog electronic
> circuit simulator
> ..
These sorts of programs were originally written in languages
other than FORTRAN.
Some even predate any implementation of FORTRAN.
nm...@cam.ac.uk
Eh? Some were, some weren't. Fortran was THE implementation language
of choice for portable scientific applications (including statistical
and engineering ones) between 1960 and 1980. It was pretty heavily
used for non-portable ones, too.
>Some even predate any implementation of FORTRAN.
The number of such things written before 1956 (which is when I believe
the first implementation of Fortran was released) was pretty damn
small, and the applications would be regarded as minor utilities
nowadays. I doubt that any of them have any direct descendants
still in use.
Regards,
Nick Maclaren.
Dr Ivan D. Reid
CERN's Physics Analysis Workstation (PAW) was originally written in Fortran
I believe, as was CERNLIB, the underlying libraries. It was widely used
in HEP and elsewhere, but is now largely superseded by ROOT which is C++
based. (There was a PAW++ which may have been C++ based too, I don't
recall the details.)
--
Ivan Reid, School of Engineering & Design, _____________ CMS Collaboration,
Brunel University. Ivan.Reid@[brunel.ac.uk|cern.ch] Room 40-1-B12, CERN
KotPT -- "for stupidity above and beyond the call of duty".
robin
That wasn't what the OP asked.
In any case, ALGOL was the language of choice for UK and Europe.
PL/I also was the language of choice.
> >Some even predate any implementation of FORTRAN.
>
> The number of such things written before 1956 (which is when I believe
> the first implementation of Fortran was released) was pretty damn
> small,
Programs for solving engineering and numerical problems
were written and solved on Pilot ACE prior to 1955,
which date precedes any implementation of FORTRAN.
These programs were used extensively on DEUCE
and ACE.
Much of the numerical work was pioneered by J. H. Wilkinson.
For DEUCE, the published programs numbered in excess
of 1,000 in all fields.
> and the applications would be regarded as minor utilities
> nowadays.
Nonsense. They were used for structural design, aircraft design,
nuclear design, and the like.
> I doubt that any of them have any direct descendants
> still in use.
That wasn't what the OP asked.
In any case, those applications were rewritten in
some HLL such as ALGOL, FORTRAN, and PL/I.
Addressing the OP's question,
"I'm writing a paper on software programs originally written in Fortran ..."
what I said was that many of those applications were
originally written in languages other than FORTRAN.
You might take a look at CACM algorithms.
The early ones (1960s) were re-inventing wheels:
1. Integration
2. Rootfinder
3. Polynomial roots
4. Rootfinding, bisection method
5,6. Bessel functions
7. GCD
8. Euler
9. Runge-Kutta
10. Chebychev
11.Polynomial evaluation
12. Laguerre polynomial evaluation
13. Legebndre polynomial evaluation
14. Complex exponential integral
15. Crout's method for solving simultaneous equations
and so on.
These elementary procedures had been in use
well before 1955.
Of course, much more elaborate programs of the
kind that I mentioned were in use by 1955,
and continued to be developed through 1965.
These dates precede all those dates cited by the OP,
the earliest of which was 1970.
nm...@cam.ac.uk
robin <rob...@bigpond.com> wrote:
>> >
>> >These sorts of programs were originally written in languages
>> >other than FORTRAN.
>>
>> Eh? Some were, some weren't. Fortran was THE implementation language
>> of choice for portable scientific applications (including statistical
>> and engineering ones) between 1960 and 1980. It was pretty heavily
>> used for non-portable ones, too.
>
>That wasn't what the OP asked.
No, it wasn't. I was responding to your categorical statement, which
you may have made rather more absolute than you intended.
>In any case, ALGOL was the language of choice for UK and Europe.
Sigh. Do look at my Email address. No, it wasn't. Many people
used ALGOL 60, many used Fortran, and many used other languages.
In the 1960s (but no later), many people used Autocode, but
abandoned it because it wasn't portable. Most of Rothamstead's
and RAL's codes were in Fortran between those dates, to name
only two important sites.
>PL/I also was the language of choice.
Not for portable applications. In the statistical field, there were
many dozens of Fortran ones, and the very few PL/I ones (SAS etc.)
were nor very portable or even specific to a single system. That
also seemed to be true in engineering (ICES etc.) and in the other
'scientific' areas I had some contact with.
>> and the applications would be regarded as minor utilities
>> nowadays.
>
>Nonsense. They were used for structural design, aircraft design,
>nuclear design, and the like.
Those computers had 1-4 KB of memory. Yes, they were used for those
purposes (I did so on a Mercury Meteor), but the applications were
what would be regarded as minor utilities nowadays.
>> >Some even predate any implementation of FORTRAN.
>>
>> The number of such things written before 1956 (which is when I believe
>> the first implementation of Fortran was released) was pretty damn
>> small,
>
>Programs for solving engineering and numerical problems
>were written and solved on Pilot ACE prior to 1955,
>which date precedes any implementation of FORTRAN.
>These programs were used extensively on DEUCE
>and ACE.
>Much of the numerical work was pioneered by J. H. Wilkinson.
>For DEUCE, the published programs numbered in excess
>of 1,000 in all fields.
Considering that the first DEUCE computer was delivered in May 1955,
I doubt that many of those predated Fortran.
Regards,
Nick Maclaren.
Peter
In my field, antenna engineering, there are many Fortran codes in
general use. Here are just a few...
Ohio State University has continuously developed a suite of antenna
programs in Fortran since the 1970s, of which perhaps the best known
is the Basic Scattering Code (http://esl.eng.ohio-state.edu/~rjm/codes/
infnzbsc.htm).
The Danish company Ticra sells a number of commercial codes written in
Fortran which feature object orientation and full GUI interfaces.
These codes are standards in the satellite antenna industry, and have
also been continuously developed and improved for several decades.
The most recent versions are written in Fortran 95, I believe. Major
codes include GRASP, CHAMP, and POS. More info at www.ticra.com
--Peter
Brian Salter-Duke
> In article <NCygm.11240$ze1....@news-server.bigpond.net.au>,
> robin <rob...@bigpond.com> wrote:
>>> >
>>> >These sorts of programs were originally written in languages
>>> >other than FORTRAN.
>>>
>>> Eh? Some were, some weren't. Fortran was THE implementation language
>>> of choice for portable scientific applications (including statistical
>>> and engineering ones) between 1960 and 1980. It was pretty heavily
>>> used for non-portable ones, too.
>>
>>That wasn't what the OP asked.
>
> No, it wasn't. I was responding to your categorical statement, which
> you may have made rather more absolute than you intended.
>
>>In any case, ALGOL was the language of choice for UK and Europe.
>
> Sigh. Do look at my Email address. No, it wasn't. Many people
> used ALGOL 60, many used Fortran, and many used other languages.
> In the 1960s (but no later), many people used Autocode, but
> abandoned it because it wasn't portable. Most of Rothamstead's
> and RAL's codes were in Fortran between those dates, to name
> only two important sites.
I agree with this. I think my experience in UK was typical. I learnt
Mercury Autocode on the Oxford Mercury in 1960. I did some stuff in
Elliot Autocode at Reading Uni in 1963. I learnt ALGOL at Newcastle Uni
in 1964, but slowly moved over to FORTRAN particurly when I moved to
Lancaster in 1966 because of the growing bulk of FORTRAN code in what is
now called computational quantum chemistry.
Fortran codes in this field were or are:-
ATMOL
Hondo
Gaussian (from Gaussian72 in 1972)
GAMESS(US)
GAMESS(UK)
CADPAC
I have forgotten the names of some other very early US codes. There is
also the MOPAC semi-empirical code and the many molecular mechanics
codes starting I think with MM2, but htere might I suppose have been a
MM1.
--
Brian Salter-Duke Melbourne, Australia
My real address is b_duke(AT)bigpond(DOT)net(DOT)au
Use this for reply or followup
"A good programmer can write Fortran in any language"
Robert Orban
jason.li...@gmail.com says...
>
>
>I'm writing a paper on software programs originally written in Fortran
>(or fortran variant) that was (or is being) used for science and
>engineering. I was wondering if someone could help me with
>my list..? Here is what I found so far:
>
>1. HEC-RAS U.S. Army Corps of Engineers 1991 River Channel Water
>Surface Profile
>2. MODFLOW USGS 1983 3D Ground-water flow model
>3. ANSYS Dr. John Swanson 1970 finite element analysis,
structural
>analysis
>4. HFPF U.S. EPA Hydrology & Water Quality Water Bodies
>5. EPANET Lewis Rossman, U.S. EPA 1993 Dynamic Pipe Flow
Distribution
>6. SMSIM David M. Boore, USGS 1996 Ground Motions from
Earthquakes
>7. SPICE1 Laurence Nagel, Donald Pederson 1972 Analog electronic
>circuit simulator
>..
>
>that might already exist.
--George Szentirmai's S/FILSYN filter design program: http://www.alkeng.com/.
This has been ported to Windows but the port still in Fortran (Lahey, I
believe). The program also can generate Fortran subroutines that implement
digital filters it designs.
--The classic suite of programs in the book "IEEE Programs for Digital Signal
Processing." In particular...
The Parks-McClellan-Rabiner program for designing finite impulse response
digital filters with Chebychev error criteria using the Remez Exhange
Algorithm. This is probably the single most widely used algorithm for digital
filter design to this day and has since been translated into a number of
languages other than Fortran. These days, it is mainly found as a routine
available inside a larger filter design package.
erga
> I'm sorry for not being more specific and for asking a misleading
> question. I'm trying to compare Fortran to other early languages that
> are still in use today. I want to provide some examples to show why
> the languages are used and for what purpose. So I am making a list of
> examples that demonstrate an idea, not a comprehensive list of every
> program. I've read that Cobol is good for business, Lisp for AI, and
> Fortran for science/engineering. I want to show examples that
> illustrate that (or maybe something different). I will make sure that
> my audience understands that this list is a sampling only.
>
> My current list of fortran programs so far is mainly confined to civil
> engineering tools (except for Spice1) - I am a civil engineer, and I
> am familiar with these. I thought if I ask this question here, I would
> get a more well-rounded response that would be more representative.
Applications in animal breeding / statistical quantitative genetics
are common since at least the 60's (I am too young to remember :-) )
and always in use. These involve solving huge mixed models (up to some
million equations) with sparse matrices and the practical use is to
estimate genetic values of animals (cattle, swine, poultry...). There
are also applications used in fancier stuff such as gene mapping by
linkage analysis; for example Simwalk2.
Of course there are thousands of self-made and lab internal programs.
A few that you can find on the web:
Blupf90, remlf90: http://nce.ads.uga.edu/~ignacy/newprograms.html
Wombat: http://agbu.une.edu.au/~kmeyer/homepage.html
ASREML: http://www.vsni.co.uk/software/asreml/
Simwalk2: http://watson.hgen.pitt.edu/docs/simwalk2.html
--
robin
> In article <pan.2009.08.10....@none.net>,
> none <no...@none.net> wrote:
>
> [ A good summary ]
>
> >early to late 1980s - Pascal began to take over Fortran share.
>
> Grrk. Not really. It was mainly kiddies and computer scientists.
> Pascal was dire for practical programming, especially for programs
> that used arrays seriously.
The problem was that arrays were all static.
That and the fact that it handled strings poorly.
> >No one doubts that there are better languages than Fortran for many tasks.
> >Personally I felt that Algol 68 was a worthy successor, but it didn't take
> >off - & I was using iut in the late 1980s.
>
> Yes, I agree. I was using it from about 1970 on. However, Fortran
> has now lifted several of Algol 68's innovative ideas ....
Even from Algol 60 and PL/I.
Terence
Brian Salter-Duke wrote:
> I agree with this. I think my experience in UK was typical. I learnt
> Mercury Autocode on the Oxford Mercury in 1960. I did some stuff in
> Elliot Autocode at Reading Uni in 1963. I learnt ALGOL at Newcastle Uni
> in 1964, but slowly moved over to FORTRAN particurly when I moved to
> Lancaster in 1966 because of the growing bulk of FORTRAN code in what is
> now called computational quantum chemistry.
>
> Fortran codes in this field were or are:-
>
> ATMOL
> Hondo
> Gaussian (from Gaussian72 in 1972)
> GAMESS(US)
> GAMESS(UK)
> CADPAC
>
> I have forgotten the names of some other very early US codes. There is
> also the MOPAC semi-empirical code and the many molecular mechanics
> codes starting I think with MM2, but htere might I suppose have been a
> MM1.
>
> Brian Salter-Duke Melbourne, Australia
> My real address is b_duke(AT)bigpond(DOT)net(DOT)au
> Use this for reply or followup
> "A good programmer can write Fortran in any language"
I share many parallels with Brian, in the UK phase:
Mercury Autocode for Blue Streak guidance at De Havilland in 1955
before later switching to Fortran, (and also ending up in Australia!)
Adding to the list.
My first post-introductory working day in IBM in 1961 I was given a
Fortran manual and told to help Balfour Beatty write, in Fortran, the
code that designed electricity transmission towers, from the top down,
(starting with cable angles, spans covered and cable weight/foot to
give weights and torques) and generating a mechano-type parts length
list of girder piece sizes and forms to build the specific tower, then
on getting to the soil line, design the foundations. I recognise these
designs now, world-wide.
Later, Shell UK wrote their accounting system in Fortran and ran it
at the IBM Service Bureau (1961). Since I worked for Shell from 1966
on, I know almost all their code for simulations and engineering was
in Fortran (or some I wrote in PL/1, whe we adopted to bring in Cobol
programmers from the cold, into a single company-wide programming
discipline)
robin
> In article <NCygm.11240$ze1....@news-server.bigpond.net.au>,
> robin <rob...@bigpond.com> wrote:
> >> >
> >> >These sorts of programs were originally written in languages
> >> >other than FORTRAN.
> >>
> >> Eh? Some were, some weren't. Fortran was THE implementation language
> >> of choice for portable scientific applications (including statistical
> >> and engineering ones) between 1960 and 1980. It was pretty heavily
> >> used for non-portable ones, too.
> >
> >That wasn't what the OP asked.
>
> No, it wasn't.
That's right. It wasn't.
> I was responding to your categorical statement, which
> you may have made rather more absolute than you intended.
Let me say it again. Those sorts of programs
were written originally in languages other than Fortran.
Do you think, for instance, that the early algorithms that appeared
in CACM were the first instances of those algorithms?
Of course not; they were written in machine code or some other
computer language five to ten years prior. Some were implementations of
algorithms that had been known for decades and solved by hand
computation.
> >In any case, ALGOL was the language of choice for UK and Europe.
>
> Sigh. Do look at my Email address. No, it wasn't. Many people
> used ALGOL 60, many used Fortran, and many used other languages.
And some wrote their own Algol compiler.
Not language of choice?
You should note that early British and European
computers in the public eye used mainly Algol and languages
other than FORTRAN. Even in the US that was the case.
> In the 1960s (but no later), many people used Autocode, but
> abandoned it because it wasn't portable.
It was portable to other machines of the same kind.
But autocodes were generally primitive, and were
abandoned not because they weren't portable, but because
some better language came along.
> Most of Rothamstead's
Rothamsted
> and RAL's codes were in Fortran between those dates, to name
> only two important sites.
> >PL/I also was the language of choice.
>
> Not for portable applications.
Yes, for portable applications. Also for general algorithm design,
like Algol.
You're forgetting that IBM's PL/I SSP was able to run on
any PL/I compiler, including their DOS subset.
You are also forgetting that FORTRAN codes were no more
portable than a log. Apart from the need to change the code
to run on another model of the same machine, porting to
another manufacturer involved removing/changing extensions,
in particular changing the way subroutines handled arrays,
changing any character handling, and the like.
> In the statistical field, there were
> many dozens of Fortran ones, and the very few PL/I ones (SAS etc.)
> were nor very portable or even specific to a single system. That
> also seemed to be true in engineering (ICES etc.) and in the other
> 'scientific' areas I had some contact with.
>
> >> and the applications would be regarded as minor utilities
> >> nowadays.
> >
> >Nonsense. They were used for structural design, aircraft design,
> >nuclear design, and the like.
>
> Those computers had 1-4 KB of memory.
The amount of memory a system had (or has)
has very little to do with the complexity of algorithn that can be written.
In comparison with current machines, memories then were small.
But codes were compact.
> Yes, they were used for those
> purposes (I did so on a Mercury Meteor), but the applications were
> what would be regarded as minor utilities nowadays.
I repeat that you are wrong.
Programs for solving engineering problems used OO programming.
In any case I know of at least one problem required the solving
of more than 1,000 simultaneous equations.
One compiler occupied 8KB, such was the compact code
of some early systems.-
Important sites in the UK included UKAEA,
Wharton (Atomic Power Division of English Electric),
Central Electricity Generating Board (CEGB),
RAE at Farnborough; British Aero Engine Co.;
BP, Ministry of Agriculture, Fisheries and Food,
and of course National Physical Laboratory --
to mention a few of the sites where DEUCE was installed.
Why were those important? They were are the forefront
of engineering and scientific development. Pilot ACE
was one of the fastest computers (if not the fastest) when it
was introduced, and DEUCE (re-engineered Pilot ACE)
had the same clock speed as Pilot ACE.*)
> >> >Some even predate any implementation of FORTRAN.
> >>
> >> The number of such things written before 1956 (which is when I believe
> >> the first implementation of Fortran was released) was pretty damn
> >> small,
> >
> >Programs for solving engineering and numerical problems
> >were written and solved on Pilot ACE prior to 1955,
> >which date precedes any implementation of FORTRAN.
> >These programs were used extensively on DEUCE
> >and ACE.
> >Much of the numerical work was pioneered by J. H. Wilkinson.
> >For DEUCE, the published programs numbered in excess
> >of 1,000 in all fields.
>
> Considering that the first DEUCE computer was delivered in May 1955,
> I doubt that many of those predated Fortran.
That's irrelevant, as you well know (and I didn't say that anyway).
The programs were originally written for Pilot ACE in the early 1950s
(i.e., pre 1955) and continued on DEUCE.
The software for Pilot ACE was developed from the
late 1940s.
All of that predates FORTRAN, does it not!
As for DEUCE, the first machines (plural) were
delivered in 1955, and came with a large library
of published code inherited from Pilot ACE.
____________
* but faster because it has some better instructions
and doubled I/O capacity.
nm...@cam.ac.uk
robin <rob...@bigpond.com> wrote:
>
>Let me say it again. Those sorts of programs
>were written originally in languages other than Fortran.
>
>Do you think, for instance, that the early algorithms that appeared
>in CACM were the first instances of those algorithms?
>Of course not; they were written in machine code or some other
>computer language five to ten years prior. Some were implementations of
>algorithms that had been known for decades and solved by hand
>computation.
Some where - some weren't. I know or knew some of those people, and
what they developed in, though I have been producing such algorithms
myself only since the early 1970s.
>You should note that early British and European
>computers in the public eye used mainly Algol and languages
>other than FORTRAN. Even in the US that was the case.
Oh, God. The clarity that viewing objects from a remote distance
brings ....
English Electric KDF9 and System 4. Ferranti Atlas. ICL 1900 series.
To name but a few. All of those had Fortran compilers that were
heavily used, sometimes as the primary language.
>You are also forgetting that FORTRAN codes were no more
>portable than a log. Apart from the need to change the code
>to run on another model of the same machine, porting to
>another manufacturer involved removing/changing extensions,
>in particular changing the way subroutines handled arrays,
>changing any character handling, and the like.
Some of that was true in Fortran II, but it wasn't the case in
Fortran 66 (witness GLIM, Genstat, Clustan and hundreds of other
highly portable Fortran codes). But what changes to subroutine
array handling do you think were necessary?
>> Those computers had 1-4 KB of memory.
>
>The amount of memory a system had (or has)
>has very little to do with the complexity of algorithn that can be written.
Sigh. A lot of algorithms need more than 4,000 mathematical symbols
to express. There just ain't no way that they could be run on those
machines, because the latter's MTBF wasn't high enough to run them
in 'large machine emulation mode'.
>In comparison with current machines, memories then were small.
>But codes were compact.
Patent 123456789: A method of compiling an algorithmic description
using N symbols into one that uses only N/2 of the same symbols.
Regards,
Nick Maclaren.
Terence
me right back to the events, smells and sounds. (I took some pilot
instruction in 1946 on the BoB Shoreham fighter fields).
In the same way, the Robin-Nick interchange made me feel in the middle
again, in the computer wars of the fifties and sixties. As an
apprentice in 1951-52 at Blackburn Aircraft, UK, I helped an attempt
to build a nixie-tube readout, twin triode storage, teletype I/O
computer to measure 200 strain gauges to find why tails fell off
aeroplanes if they flew the polar routes (metal fatigue). This
eventiually got me into Ferranti Mercury usage for DeHavilland on
missiles, then IBM UK and so Fortran.
I was aware of all the IBM computers and wired computing engines, many
ICL, NCR and Burroghs units, Atlas at Manchester and so on - glorious
days!,
The names brought up by the above duo brought back clear images of
places and events and even codes. I found myself nodding here and
there but shaking my head in others, on both sides.
When older, long-ago dates are somwhat meanless without recourse to
diaries, so the "truth" is often only local. Hasn't anybody wrtten a
really thick book on the history of computing (with nods to Pascal,
Babage, Lady Ada Lovelace et al?).
It would be so easy to have such a reference book resolve these
opinions.
By the way the biggest problem with some old codes was the matrix
storage order assumed (row first or column first; the IBM version BMD
codes invert modern usage). A common search trick was to treat a 2D or
3D matrix alternatively as a one-dimensional string of values. Such a
trick wouldn't work under some implementations of Fortran in the
fifties. Other problems were Integer*1 and general lack of character
equivalents, and so how many characters went into a floating point
storage unit? 4, 5 or 6?
Michel Oui
> I'm writing a paper on software programs originally written in Fortran
> (or fortran variant) that was (or is being) used for science and
> engineering. I was wondering if someone could help me with
> my list..? Here is what I found so far:
>
>
You've got a typo in #4, it should be:
HSPF - Hydrological Simulation Program--Fortran
robin
news:h68kbu$ft4$1...@smaug.linux.pwf.cam.ac.uk...
> Some of that was true in Fortran II, but it wasn't the case in
> Fortran 66 (witness GLIM, Genstat, Clustan and hundreds of other
> highly portable Fortran codes)
They weren't "highly portable".
As I said, they had to be altered to run even on different models of the
same
machine!
Genstat, for example, had different versions for mainframes, minis, unix,
DEC, Honeywell, IBM, ICL, Pr1me, Sun, Unisys, PCs.
That was the reason that PL/I and Algol were far superior --
they used dynamic storage, not rigid array bounds as in FORTRAN.
Gordon Sande
> <nm...@cam.ac.uk> wrote in message
> news:h68kbu$ft4$1...@smaug.linux.pwf.cam.ac.uk...
>
>> Some of that was true in Fortran II, but it wasn't the case in
>> Fortran 66 (witness GLIM, Genstat, Clustan and hundreds of other
>> highly portable Fortran codes)
>
> They weren't "highly portable".
> As I said, they had to be altered to run even on different models of the
> same
> machine!
The more professional ones had preprocessors to insert the dimensions
for various models. And the preprocessors were indeed portable as they
were rather small programs.
Look at the RxMACH type suboutines in netlib to see the general style.
There were lots of examples of overly clever programs that would not even
survive a minor change in version number but such foolishness got weeded
out after a couple of sessions.
> Genstat, for example, had different versions for mainframes, minis, unix,
> DEC, Honeywell, IBM, ICL, Pr1me, Sun, Unisys, PCs.
>
> That was the reason that PL/I and Algol were far superior --
> they used dynamic storage, not rigid array bounds as in FORTRAN.
The dynamic storeage solved one problem and left multiple others unsolved.
robin
news:2009082815241016807-gsande@worldnetattnet...
> On 2009-08-28 11:39:51 -0300, "robin" <rob...@bigpond.com> said:
>
>> <nm...@cam.ac.uk> wrote in message
>> news:h68kbu$ft4$1...@smaug.linux.pwf.cam.ac.uk...
>>
>>> Some of that was true in Fortran II, but it wasn't the case in
>>> Fortran 66 (witness GLIM, Genstat, Clustan and hundreds of other
>>> highly portable Fortran codes)
>>
>> They weren't "highly portable".
>> As I said, they had to be altered to run even on different models of the
>> same
>> machine!
>
> The more professional ones had preprocessors to insert the dimensions
> for various models. And the preprocessors were indeed portable as they
> were rather small programs.
>
> Look at the RxMACH type suboutines in netlib to see the general style.
I'm aware of those crude procedures, which are unnecessary in PL/I.
> There were lots of examples of overly clever programs that would not even
> survive a minor change in version number but such foolishness got weeded
> out after a couple of sessions.
>
>> Genstat, for example, had different versions for mainframes, minis, unix,
>> DEC, Honeywell, IBM, ICL, Pr1me, Sun, Unisys, PCs.
>>
>> That was the reason that PL/I and Algol were far superior --
>> they used dynamic storage, not rigid array bounds as in FORTRAN.
>
> The dynamic storeage solved one problem and left multiple others unsolved.
Such as?
Nik Dattani
I would have asked the author directly, but the only information I have about the author is that their name is "Jason" (no last name, and no email address available?).
FortranFan
> Sorry for replying to such an old thread, but I wonder if this paper ever got written?
>
> I would have asked the author directly, but the only information I have about the author is that their name is "Jason" (no last name, and no email address available?).
If your interest with the paper mentioned in the original post is with some list of applications in Fortran, as explained in the replies that is too arduous a task to compile. However if you seek a broad sample of Fortran codes and you are content to limit the list arbitrarily e.g., to what is available on GitHub, take a look at the link below for a remarkable list maintained by @Beliasvsky:
https://github.com/Beliavsky/Fortran-code-on-GitHub
Jeff Ryman
> > I'm sorry for not being more specific and for asking a misleading
> > question. I'm trying to compare Fortran to other early languages that
> > are still in use today. I want to provide some examples to show why
> > the languages are used and for what purpose. So I am making a list of
> > examples that demonstrate an idea, not a comprehensive list of every
> > program. I've read that Cobol is good for business, Lisp for AI, and
> > Fortran for science/engineering. I want to show examples that
> > illustrate that (or maybe something different). I will make sure that
> > my audience understands that this list is a sampling only.
> >
> > My current list of fortran programs so far is mainly confined to civil
> > engineering tools (except for Spice1) - I am a civil engineer, and I
> > am familiar with these. I thought if I ask this question here, I would
> > get a more well-rounded response that would be more representative.
> _Small_ sampling would include
> PDQ/HARMONY -- neutron diffusion/isotope depletion combination
> RELAP -- LWR reactor safety (LOCA, etc., ...) analysis
> KENO -- Monte Carlo particle transport
> ANISN, DOT3,5 -- Discrete ordinate particle transport
> QAD-CGGP2 -- point kernel shielding
> ORIGEN -- Isotope Generation and Depletion
> The Radiation Safety Information Center (RSIC) at ORNL (Oak Ridge
> National Laboratory) is the distribution center for NRC-supported codes
> for computations in the field that will have abstracts of these and many
> more.
> EPRI (formerly known as Electric Power Research Institute) has developed
> many software packages for a wide range of problems for the electric
> utilities from fossil-fired plant thermal performance monitoring (PMW)
> to transmission/distribution network modeling.
> Don't know particular codes but at one time many of the petroleum
> exploration/production companies geophysical modeling of production
> fields were Fortran on CDC/Cray as would run into those guys at a couple
> of the timeshare facilities way back when...they were much more
> closed-mouth about things than anybody else I ever ran into, though,
> other than the most general of generalities so information in that area
> is probably going to be much harder to find.
> --
Most of the codes in the SCALE code system (you mentioned KENO and ORIGEN),
although they are being converted to C++ for reasons unfathomable to me. Also
the AMPX nuclear cross section processing system (many codes within). Then
from Los Alamos Lab MCNP and MCNPX for Monte Carlo Radiation Transport,
and NJOY for nuclear cross section processing, although the latest version is
now completely in C++.
robin vowels
Five years of programming ion Pilit ACE did not amount to
anything?
Many of the programs already in use on Pilot ACE
were re-written for DEUCE. The instructions sets
were similar.