dph@lanl.gov (David Huelsbeck) (02/25/89)
From article <23612@beta.lanl.gov>, by dd@beta.lanl.gov (Dan Davison): > In article <27220@sgi.SGI.COM>, jmw@rand.SGI.COM (James M Winget (Ext 3654)) writes: >> I'm looking for SW tools for analysis of large Fortran Programs. >> (What is the current status of TOOLPACK ???) > > > A Toolpack mailing list is being maintained by dph@lanl.gov. He has > been working on Version 2, and has been doing a lot with it. Try > e-mailing to him to get on the list and the latest info. Well, it seems a lot of you have tried and succeeded. My mailbox runneth over. If you sent me mail or you're thinking about sending me mail hang on until the first part of next week. I'm preparing a blanket response that will, I hope, answer all of your questions. Until then, if you really can't wait to find out more about Toolpack/1 check out the article "Implementation of Toolpack/1 on a Graphics Workstation" in Software P & E 1988 or the original IEEE Trans. on Software Eng. article by Leon J. Osterweil, this was '79 or '80. There was another SP&E article circa 1980 but I couldn't find it. I know that these aren't wonderful references but complete references will be included in my follow up early next week. Thanks for you patience, -dph
dph@lanl.gov (David Huelsbeck) (03/03/89)
Well, I had in mind putting together something a bit more formal than this. However, it seems the time needed is going to be hard to find for a while. So here's a brief summary of what is Toolpack/1, how can you get it and what's its status. Toolpack/1 Release 1 was the result of an international collaborative research project. It was intended to be a tightly intergrated set of tools to form an environment for the development and maintenance of "medium-sized mathematical software projects written in standard-conforming Fortran/77." A secondary goal was the investigation of extensible software development environments using integrated tool suites. As a result of its research nature and the funding from US DOE, NSF and the Science and Engineering Research Council of the UK it was public domain. NAG agreed to distribute it for a nominal fee. Something like 400 sites received the original distribution. After some collective experience had been gained, some bugs and deficiencies uncovered, etc, NAG took the initiative and spruced it up some. I am told they did quite a bit of wholesale recoding of their own as well as collecting and packaging the contributions of others. The result was Toolpack/1 Release 2 which is currently at version 2.5. In the spirit of the original project this is still in the public domain. From what I understand it is used somewhat more widely in Europe than it is in the US. I personally know of a dozen or so sights in the US and a couple in Canada that are using it now. Speaking from my personal experience with it, I'd say it's quite good considering the wide range of things it attempts to do. It includes a static analyzer, a code polisher (very sophisticated pretty-printer), a restructurer (despaghettizer) and much more. It's also supposed to be portable. While I haven't yet tried to port it there are already versions for virtually all major hardware/OS combinations. It has two important weaknesses though. It won't handle very large (~100,000 line) in one bite. This limits the usefulness of the static analyzer and makes everything else less convenient. It also adheres *VERY* strictly to the standard. This may be a feature if you're looking to write portable code, but it can be a serious problem if you use some special dialect. I've been discussing means of fixing the former problem but the later seems pretty sticky. At least one vendor that I know of is working on a version for their specific dialect, but their work is not likely to be of aid to anybody else. So now that you're interested how do you get a copy of your very own? IFF you have ftp capability you can ftp a copy from lambda.lanl.gov in the directory ~ftp/pub/toolpack. There's a README file there that will tell you what the other files are. If you don't have ftp access don't write me asking for a tape. NAG provides that service for a small fee. To get in touch with NAG mail to: Numerical Algorithms Group, Inc 1101 31st Street, Suite 100 Downers Grove, IL 60515-1263 USA or for those in Europe Numerical Algorithms Group, Ltd NAG Central Office Mayfield House 265 Banbury Road Oxford OX2 7DE UK If their fees are too high for you I guess you'll have to get creative. Finally, as was mentioned in the original posting, there is a mailing list for discussing issues related to Toolpack/1. To get on the list send mail to me (dph@lanl.gov) and try to give me as much mail routing information as you can. I was at one time requesting that new subscribers send a short blurb on who they are and what their interest in Toolpack is but you may consider this optional. An archive of the list discussion is also there for ftp in the toolpack source directory. Everyone who sent me mail requesting addition to the list has been added. However I'm not sure that Leo C. Noordhuizen of the Phillips CAD Centre in Eindhoven, NL is getting my mail. If you read this, please send me some more mail with as much routing information as you can give me. Oh, I almost forgot. I promised a bibliography in my original posting. Osterweil, L. J., "Toolpack - An Experimental Software Development Environment Research Project," IEEE Transactions on Software Engineering, vol. SE-9, no. 6, pp. 673-685, 1983 Broom, M. A. and Hopkins, T. R., "The Implementation and Use of Toolpack/1 on a Graphics Workstation," Software Practice and Experience, vol 17, no 9, pp. 561-573, 1987
2011_552@uwovax.uwo.ca (Terry Gaetz (UWO Astronomy); Bitnet: gaetz@uwovax) (03/04/89)
In article <9973@lanl.gov>, dph@lanl.gov (David Huelsbeck) writes: [...] > So now that you're interested how do you get a copy of your very own? > IFF you have ftp capability you can ftp a copy from lambda.lanl.gov > in the directory ~ftp/pub/toolpack. There's a README file there that -- Does anyone know the numeric address for lambda.lanl.gov? Our site doesn't have a name server yet. Thanks in advance! -- Terry Gaetz gaetz@uwovax.uwo.ca | gaetz@uwovax.BITNET
2011_552@uwovax.uwo.ca (Terry Gaetz (UWO Astronomy); Bitnet: gaetz@uwovax) (03/04/89)
In article <1785@uwovax.uwo.ca>, 2011_552@uwovax.uwo.ca (Terry Gaetz (UWO Astronomy); Bitnet: gaetz@uwovax) writes: [...] > Does anyone know the numeric address for lambda.lanl.gov? Our site doesn't > have a name server yet. Thanks in advance! > -- I have already received 3 responses; thanks to all for their prompt replies! (The answer is 128.165.4.10, and it works.) -- Terry Gaetz gaetz@uwovax.uwo.ca | gaetz@uwovax.BITNET