clemh@nvuxr.UUCP (C Hergenhan) (08/03/86)
I need to ship some large files around soon and am looking for a way to compress the data. Some other info 1) the chunks are around 50k bytes each so a "learning" algorithm has to pay back on this size, 2) many of the bytes contain two 4-bit BCD digits, i.e, there are only 9 patterns in the 4-bits, 3) the chunks consist of records which have some repeated strings in them, e.g., the time and date (expressed in BCD digits) is in each record and will probably be the same for half the records in the file, 4) there will be lots of systems involved so having a permanent (and correct) translation table at each end will be hard to maintain, i'd prefer the compressed chunk have enough info within it for decompressing. I don't know much about compression techniques so I'm not sure the above items are the key points to a solution. any suggestions on an algorithm (and accompanying C source to try on my data) are welcome. thanks clem hergenhan, 201-747-4700, ...!nvuxr!clemh