From owner-metabolic-reg@net.bio.net Fri Sep 10 19:35:00 1999 Path: biosci!newsfeed.stanford.edu!newsfeed.berkeley.edu!dispose.news.demon.net!demon!news.demon.co.uk!demon!fssc.demon.co.uk!not-for-mail From: "Herbert M Sauro" Newsgroups: bionet.metabolic-reg Subject: Reasons...... Date: Fri, 10 Sep 1999 20:30:30 +0100 Message-ID: <936995329.9034.0.nnrp-12.c2de3927@news.demon.co.uk> References: <936947725.17873.0.nnrp-04.c2de3927@news.demon.co.uk> NNTP-Posting-Host: fssc.demon.co.uk X-NNTP-Posting-Host: fssc.demon.co.uk:194.222.57.39 X-Trace: news.demon.co.uk 936995329 nnrp-12:9034 NO-IDENT fssc.demon.co.uk:194.222.57.39 X-Complaints-To: abuse@demon.net X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.00.2014.211 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2014.211 Lines: 107 I think some clarification is required on the role of pmb files since it wasn't clear perhaps why it was announced. Athel Cornish-Bowden wrote in message news:v03007801b3fed8bb3cc4@[193.50.234.80]... > Herbert Sauro wrote: > > >Portable Metabolic Binary Standard (pmb files) > > > >This message is to announce the availability of a new portable metabolic >> .........etc > >being considered along side pmb and details will be posted in a later > >announcement. > > What is the advantage of making it a binary format (and hence "not readable > by humans")? > > Most of those of us who follow this list are humans and some of the > advantages of making things readable by humans are obvious. Another > advantage, even with things to be read by machine is that text files > generate fewer problems for moving data between operating systems. ****** I hope pmb files are platform independent, there are no issues for example concerning the type of line feed used. Just this very day at work I had to manually edit in binary some source code we obtained from a Unix platform because the debugger on our development platform went crazy when it found single LFs were terminating lines (Windows uses CRLF), now that took a while to work out the cause! > > What are the compensating disadvantages? > > For large data files (e.g. graphics files, word-processor files, etc.) I > can think of some advantages in using a binary format, such as > > 1. Reading and writing are faster; > 2. For a commercial developer using a binary format makes life more > difficult for competitors, pirates etc.; > 3. A binary file may contain less wasted space. > > However, > > 1. For the sort of thing Herbert is talking about, I would be surprised if > the files got so big that one would notice any difference in speed; ******* You're right, nothing to do with speed. > 2. I wouldn't have thought that making life difficult for competitors was > an issue here; ******* You're right, it's nothing to do with keeping competitors out, quite the opposite in fact. > 3. Looking at the sizes of text files and binary files produced by > commercial programs and containing the same information, I get the > impression (possibly wrong) that any advantage in more efficient use of > storage is likely to be trivial. > ******* You're right, there is no storage problem. > Doubtless there is some important reason for using a binary format that I > haven't thought of, but I'd be interested in knowing what it is. > ******** Indeed there is, and it goes like this. Any thing which humans can meddle with will invariable end up containing errors, errors like spelling mistakes, syntactic and semantic errors. If human users require readable text files then they need a parser to read and cope with all the potential errors that can arise. These are not particularly easy things to write especially for users whose main subject specialism is not computer science. However it is easy to read binary files. These are rigidly structured and because they're binary files they will 99.99% of the time be products of other computer programs rather than humans, we can therefore assume they are error free (One can add a sprinkling of error checking but it's trivial in comparison to human written text files). Also a binary file is less likely to be messed about with by humans since they won't have a clue what the contents mean. Not everyone is equipped to write text parsers but probably most scientists could write a Fortran, C or Obj Pascal program to read a binary file. The people who have developed programs like Gepasi, Mist, E-Cell, DbSolve, MetaMod etc can now easily add a pmb import/export section to their software without the tedium and very time consuming task of writing a text parser. The next version of WinScamp for example (should be available as you read this) can import and export pmb files. The hope is that users of these software packages will be able to exchange simulation models. You'll be interested to hear that there is a new language parser being written for MCA/metabolic models which can spit out these binary files aswell as do the usual numerical stuff (+ lots more). Users will be able to write they models is a human readable text form, export it as a pmb file, give it to someone else who will import it into something like Gepasi. That's the reason. > Athel .