The two I listed are very popular and very good. They are a subset of the current fortran standard. They do not allow some of the older features of fortran which are generally sources of lots of bugs, etc. and... they enforce some programming rules aimed at reducing buggy code. They won't work well for upkeeping older existing code, but they are EXCELLENT for teaching new fortran students.
I always use the GNU fortran compiler G77 (standard available in any linux distribution) . For as far as I know, this is one of the most stable compilers. If you don't have linux machines, you can always download cygwin (
I'm trying to compile some fortran code using g77 and it won't accept command line definitions like -Dvar=value.
It worked with f77. Can someone help,please?
I would like to add a comment to this thread that I think is very important!!!
I see a lot of cross talk about fortran 77 compilers. That's all fine and dandy... but... for a course teaching "new" students about fortran or for anyone learning fortran...
I HIGHLY recommend that you learn fortran 90 or 95. Fortran 77 is well over 20 years old. ( You won't see C programmers using a 20 year old version of C and you certainly wouldn't want a computer monitor design from 20 years ago when you can have one of the models available in the stores today. )
It makes no sense to go backward in time especially when modern fortran offers some VERY beneficial features and capabilities that are not in fortran 77.
The only reason for learning just F77 that I can think of is for someone wanting to do nothing more than maintain some legacy code written long ago. Of course... even for that... F90 will do the trick.
F90 and F95 ( the current standard ) are both backwardly compatible with F77. That means anyone who does not know the newer features in fortran can still continue to program with whatever style they have developed over the years.
ELF90 is available from Lahey for a very reasonable price. The full LF95 compiler from Lahey also has a respectable price if you don't want the "PRO" version. I believe that Intel has a free F95 compiler for those working on Unix ( I don't know if it is time limited or not ). F from Imagine is also still free I believe. ELF and F are not full compilers, but are subsets. However, for new programmers, they are excellent!!! As they force the programmer to develop some safer coding styles.
To cap things off:
If you are going to learn or teach Fortran, do it right!!!
Learn ( or teach ) Fortran 90/95, NOT Fortran 77!!!
Fortran NEEDS a popularity boost to continue to survive. It won't get that boost if people are only aware of the existence of obsolete versions of the language.
I know of ABSOLUTELY NO OTHER LANGUAGE that offers the power and ease of use for technical programming that fortran does. If it "fades away" do to lack of public awareness and support.... it will be a VERY sad time for engineering and technical programmers.
I think it should be pointed out that F77/F90/95 without a decent mathematical library is almost useless in many science/engineering applications. A good place to locate fortran code for a specific mathemetical function/process is
I don't think I agree with Visitor ( above ). By that logic.... practically all languages are useless in "many science/engineering applications". The point to a language is to tell the computer what you want done and in the way you want it done. Yes... there are some languages available which have such math libraries built into them. The main problem with those languages is that they aren't standardized, they are proprietary, they are much less likely to be around tomorrow, they generally only work on a limited number of systems, they cost quite a bit and they generally require anyone using a program written in that language to have already purchased that language system ( ie... you can't just give them an executable file to run ). In addition to that... they are generally slower to execute than a well written program in a standard language such as fortran. As such... these languages are often quite popular in the lab and for research. However... there use as a hardcore programming language is questionable.
Often... that will entail using some other library or code written by others ( such as those at netlib ). I personally use fortran quite a bit. I rarely use the algorithms from Netlib or Toms, etc. Not because they aren't good algorithms ( they are ), but... because I don't need them too often.
