2-apr-2011: Repaired a counting bug in the instruction. Also replaced some files in the source distribution, based on some remarks by Takahiro Ueda (see the Forum). Date of the binaries is still the same. 1-apr is too much fun.
1-apr-2011: Release of version 4.0 beta. New features like factorization of function arguments with the FactArg statement (to use the old FactArg, you should use: On OldFactArg;) and the PolyRatFun with has two arguments: a numerator and a denominator which together define a rational polynomial. For the rest it works like PolyFun. Note that you will have to pick up the new configure.ac file when you compile the sources yourself.
28-mar-2011: Few more bug fixes. One was rather complicated and may have side effects. (Has to do with wildcard vectors and sets inside dotproducts).
19-mar-2011: Many bug fixes. These sources have not yet been placed in the open source due to some technical difficulties. (mixing C and C++ code in the makefiles).
27-aug-2010: FORM RELEASED AS OPEN SOURCE! Please see the mainpage for finding the sources. Please use the forum for commentary, questions and suggestions.
27-aug-2010: Some bug fixes, but most important: the version that is also released in the open source.
4-jun-2010: New feature: -T in the command tail, or the "ON totalsize;" or #: TotalSize ON or "Totalsize ON" in the setup file makes that FORM prints after execution the maximum total space used by expressions at any moment during execution. This is the sum of in/out/hide scratch files, the sort file and the .str file. Furthermore there are quite a few bugfixes.
23-apr-2010: Repaired the problem that transform did not know about tensors. Removed a bug from the checkpoint recovery. Put in the (prefered) alternative names tointegralnotation and tosumnotation instead of explode and implode in the transform statement.
17-mar-2010: Two bugfixes. One in transform,implode and one in the empty prompt for the external command.
11-mar-2010: Bugfixes. New facility/command: Transform. Is for argument manipulations of functions with a large number of arguments. Example: f1(1,0,1,0) and
Transform,f1,encode(1,4):base=2;
will make f1(10). There are many more varieties (not only encode) (see manual).
19-feb-2010: More bugfixes. In particular in the hide system and the ReplaceLoop command (used in the color package but should be safe for that). More efficiency in some TFORM features. Replaced the Linux and Apple versions.
9-feb-2010: Bugfixes. Replaced the Linux versions. Apple will follow soon.
13-jan-2010: Added the Apple binaries (only for pentium-like architecture).
12-jan-2010: Many little bug fixes and a big one in combinations of Off Parallel; and InParallel; New options in Print statement: Print +ss; and Print +sss; Linux versions have been replaced.
9-nov-2009: Bug fix in the new solution for the interference between dummy indices and the Keep Brackets statement.
5-nov-2009: Serious bug repair. Collateral damage in recent improvements in ParFORM. Problems with dropped expressions. Also another bug with parentheses in if and while statements.
4-nov-2009: Few more bugs repaired in the Linux versions. One with more than 300 dummy indices in a single term, another with finding expressions in the index of a .sav file and a third with a deadlock in TFORM with summed $ variables.
3-nov-2009: Bug fixes in the Linux versions. Added the +ss and +sss options in the print statement. Added "n-th root of unity" as a property of symbols as in "Symbol x#6;" would be a sixth root of unity.
10-jul-2009: Added a native Windows version for 32-bit architectures.
27-jun-2009: Made new Apple/Pentium versions. Both for 64 and 32 bits architecture. TFORM still doesn't work right on Apple. Has been left out this time.
24-jun-2009: Yet some more bugs in the pattern matcher. There is much collateral damage due to the fix of 18 march. Other bug had to do with dollars in the lhs of an id-statement. That went extra wrong in TFORM.
10-jun-2009: Little bug in reading systems-independent .sav files. Wouldn't read brackets.
22-may-2009: Repaired some more problems with patterns involving functions inside functions. Another problem with modulus calculus when the size of the number was bigger that what fits inside a FORM word. And a bug that must have existed for a long time with two substitutions inside a single dotproduct involving Also (or IdOld) as in id,p1=q1+q2; al,p2=q3+q4;
Warning: In one program, after a few hours, TFORM crashes 30Not noticed this in other programs. I am working on it (JV), but it is too big for the debugger. The program gives the message "Error!!!! Illegal bracket sequence detected in PutBracketInIndex" If anyone has noticed this message in a shorter program, please send us a bug report.
10-may-2009: Repaired a very recent bug in patterns with functions inside functions. There have been some bugs there with antisymmetric functions inside other functions and there has been some collateral damage in repairing this. Other bug that was repaired: FillExpression didn't work right.
28-mar-2009: Inconsistency between manual and actual syntax for the checkpoints. Changed in the actual syntax.
25-mar-2009: Repaired bug in the handling of the PolyFun in TFORM. It was again matad that was sensitive to this.
23-mar-2009: TFORM specific bug: in Delete Storage the caches were not invalidated. This caused problems in running matad. Fixed. Also: some syntax problems with the checkpoints repaired.
18-mar-2009: Fix with respect to pattern matching involving antisymmetric functions inside function arguments. Got sometimes the sign wrong.
17-mar-2009: Repaired a severe bug in the pattern matcher of tform. Must have been the source of quite a few mysterious crashes. Also a few other minor bug fixes.
10-feb-2009: Repaired a memory leak in polynomial division. Some other small things.
26-jan-2009: First bugfix in the pattern matching of products of non-commuting and commuting functions.
25-jan-2009: Installation of version 3.3 for the various Linux/intel/opteron executables. Installation of the manual of version 3.3. A list of new features can be found here. Executables for other versions will be made available as soon as we have them.
22-dec-2008: A large number of bugfixes. There are a few new features, but they didn't make it to the manual yet. This will be fixed up soon when more new features will become available. They are still being tested. The update was only for the linux32 and the (GNU) opteron/linux64 versions.
21-may-2008: Some more bug fixes. Stage 4, dollar variables, pattern matching in products of antisymmetric functions. Replaced the linux versions for 32 and 64 bit systems.
10-may-2008: New version for 32-bits and 64-bits Linuxes. The main difference is an enormous number of bug fixes. They concern specific tform problems, but also regular problems like with the stage 4 sorting and even a bug in the arithmetic when the numbers became larger than half their assigned space. And many more bugs. The current version has a number of new features that have not yet been completely tested and hence aren't in the manual. As soon as we are happy with them, we will update and mention what to look for. One important thing is new already: we are switching to a systems independent .sav format. This could possibly mess up your existing .sav files. Hopefully this is the last time that this happens. The best way to convert is by reading the files in with an old version and printing the expressions in text format. Then read them in the new version and Save them.
2-nov-2007: New version for 32-bits and 64-bits Linuxes. Incorporates the results of the working seminar of October, attended by Jens Vollinga, Takahiro Ueda and Misha Tentyukov. Many bugfixes and a number of small extensions. Most noticeable the inclusion of the GMP library for very large numbers on systems that support it.
24-oct-2007: Changed the clock in the opteron sequential version. Includes also some bug fixes which will come available soon in the other versions (Then also GMP support for very big numbers will be included. It is currently still being tested. Don't expect too much of it though. The current routines are rather good.).
10-sep-2007: Added new version for Linux (32 and 64 bits).
Several new features. Especially for running with tform. InParallel for running many short expressions, each in separate workers rather than splitting each one up. ClearTable for restarting with tables that get too full. ArgImplode and ArgExplode for fast notation changes in harmonic sums and Multiple Zeta Values. Many bug fixes. One fix concerns a particularly nasty 'feature': if the setting of ScratchSize is greater than 2^31 on a 64 bits computer writing would fail and the program would halt. This could be tracked to an undesirable property of the system write routine (refusing to write more than 2^31 bytes in one call). From now on that should work fine, but be careful with older versions! The manual has been updated. The loadbalancing of 16-apr-2007 has been removed again. It caused more problems than it solved. To be studied better.....
16-apr-2007: Added a new version for Linux (32 and 64 bits).
Addition of the -M startup option to be used when running several instances of FORM simultaneously from minos or make (with the -j option of make). Puts the PID in the name of the temporary files to make sure they are unique.
Three bug fixes. Two in FactArg and one in the substitutions of vectors in gamma matrices (it wouldn't match if the first declared vector would be a wildcard). Bugs reported by Takahiro Ueda.
A small change in tform: To a limited extend it can now do some load balancing on things like
L F = (a+b+c+d)^10;