We very much appreciate the way PT8 manges hyphenation data. I might have mentioned that our entire language project (including publishing and literacy) will more and more store its hyphenation data in PT8 - until my own Hyphenator is ready for production.
I also love the way I can work down the wordlist, using only keyboard-shortcuts to approve or edit hyphenation information. Such mouse-free hands-can-stay-on-keyboard makes a real difference; thank you guys.
I also appreciate the inbuilt “KI” which is making ever better proposals, as our wordlist is growing. Sometimes the wordlist is showing an incomplete word, and the moment I hit CTRL+SHIFT+A the little edit-window brings up a brand-new proposal which has got the missing possible word-break. This is amazin, almost as if PT8 is having “yet another look” when I launch the edit-window. I guess that normally new proposals are only generated, when a user is opening the wordlist. So when I work a lot of new hyphenation information, I guess PT8 is learning new valid syllables and can therefore improve its proposals.
Here is my suggestion for today:
I noticed that PT8 never ever proposes any one-letter-syllables. Yet for a certain orthography we need this and we have entered hundreds on word-initial one-letter-syllables in the wordlist. Still PT8 is stubborn. And this is creating a lot of boring and stupid work for the team: Open the edit-window, move the cursor one over, enter the “=”, confirm by hitting ENTER. Hundreds of times…
I therefore propose to lift this limitation and also “accept” and therefore propose one-character-syllables if a human user has entered manually at least 36 such examples (per character).
I know that technicians are wary of such “short” items, as there might be a danger of too many false hits. But I am talking about many many words, where all the syllables are already known by PT8 and it still proposes “nothing”. Example: “akɛɛshɩɩ” PT knows already the syllables “kɛɛ” and “shɩɩ” and it has seen hundreds of word-initial “a”. So where is the risk please?
A note about our downstream workflow: Yes, in printing, you would rarely brake a line like this a=kɛɛshɩɩ. But depending on context, it might arrive to be the lesser evil, for example where other huge number-words show up nearby. We have found a way to enhance the also very nice PT8 Print Preview so that we have control over minimum number of characters for line-breaks. This is more recent than another post I made in this forum, where I still tried hacking or swapping the hyphenated_words.txt. We can now keep the full linguistic information inside PT8 at any time and still only apply typographically appropriate (according to font sizen and column width etc.) word-breaks for line-breaks. And where the layouter hits a tricky spot, he or she can look up the word in PT8 and can override inside the PDF or inside the BOOK_project-draft.SFM before printing. So yes, having one-character-syllables might sound “bad” to some of you, but just to say that “we know what we are doing”.
An alternative idea - for advanced users - would be that the PT8 hyphenation KI could make its internal list of “learnt syllables” more transparent; i.e. display it upon request. Many spell checkers allow users to see/edit the learnt internal word-list and thereby allow for checking and potentially corrections.
Again, this is not a complaint, the hyphenation tool in PT8 is very helpful already. The developers just never imagine enough how “weird” real languages and real orthographies out there in the wild can be. So please lift the one-character-limitation.
General
18
Paratext
2.1k
PTXprint
264
Paratext Lite
64
FLExTrans
17
Scripture Forge
12
Publishing Assistant
11
Platform.Bible
0
