sommar@enea.UUCP (Erland Sommarskog) (08/26/87)
This is a (non-commercial) announcement for a programme I've written. I'm posting the announcement to the groups there might where be people that are interested. (So beware if you follow-up!) FREQUENCY is a programme that reads a text and generates a frequency table for the words in the text. The basic idea is that spelling errors due to carelessness often appears just once or twice. Thus, checking through these groups will help you to spot them. Alas, the group of words with just one occurrence is about 40-50% of all words. Anyway, it is a good help to get the words separated. Compared with the famous Unix programme "spell", FREQUENCY requires much more work of the user. However, the human eye sees more. "Spell" won't complain about "witch" in a technical specification, but you perhaps would? And "spell" only works only for English. I'd assume there are similar programmes for other languages, but at least the only Swedish I know of has a very poor dictionary. And finally, "Spell" doesn't what a syllabication is. FREQUENCY does. Even if they cross page boundaries. The principle of FREQUENCY makes it almost language independent. It must know which characters that are letters. This is not obvious with a 7-bit set. It is also fond of trying to sort correctly. There is direct support for Swedish, Danish, German, English, French and Italian. I do not intend to post the code directly to the net, since all of it is written in Swedish. Only user manual is available in English. Instead those who are interested, mail me and I send you a copy. If there is a *huge* interest, I will post it to comp.os.vms. The programme is written in VAX-Pascal, thus it only runs on VMS. There is a Unix version too, although not really as fancy as V3, the current version. It does all the basic work, though. That one is written in Ada. If you are really hot, I could try to put something together for you. -- Erland Sommarskog ENEA Data, Stockholm sommar@enea.UUCP