>"J.Pietschmann" <[EMAIL PROTECTED]>
>Sorry for being unclear and short-spoken, I didn't meant to offend you.

Jeez, I was not. I think _I_ was not clear when posting the files.

>However, did you really start with an empty file in an editor and typed
>in all the pattern strings?
>There are Original Works. If someone creates a new file by typing
>stuff into an editor he creates an Original Work. Running the
>hyphenation pattern programm on a properly marked up dictionary or
>corpus is probably also creating an Original Work, with certain
>caveats (it wasn't, for example, if the dictionary was marked up
>by someone else for the sole purpose of serving as input for the
>hyphenation pattern program).
>Again, you probably used some data to derive the patterns, be it a
>corpus or a TeX file. [...]

Creating patterns for Portuguese is much simpler than with English. While
hyphenation in English is driven by etymology and is highly irregular
(hence the pattern generator procedures), in Portuguese it is dictated
solely by prosody and is highly regular. Only a few pathologic cases with
irregular pronunciation must be treated as exceptions.
So what I did was to write patterns for all possible combinations of
letters. Again, that was possible because there are lexical rules
restricting the valid combinations. When a new word enters a Portuguese
dictionary, it must conform to these rules -- for example, "basket" became
basquete, because 'k' is not part of our alphabet and words cannot end with
a 't'.
This happens with other languages as well, (e.g. Spanish and Turkish), and
this affects the size of the hyphenation files, which are much smaller. And
while the English hyphenation patterns can fail for some new strange word
that depparts statistically from the others, the Portuguese patterns work
with any word, past or future, that conforms to the current (2003 AD)
Portuguese lexical rules.

That means the only other works I used to write the files were:
      - a Portuguese grammar, listed in the file;
      - a Portuguese dictionary, to check some spellings, also referenced;
      - the FOP documentation and a file describing the TEX algorithm (I
thought it wasn't necessary to reference that).

>In this file you generously transferred your copyright to the ASF (this
>fact crept into my brain only after I committed the file).

Actually I did this in my spare time, and only some months later I came to
use it in a company project that used FOP (with my influence, of course).
So the file is not property of Petrobrás, it is mine. But my employee
wouldn't care anyway. Petrobrás' business is oil, not software, we only
keep to us software that carries proprietary technology (our petrochemical
process simulator, for example).

I am donating the hyphenation file to the ASF, and although it would be
nice to keep the copyright, I think that would hamper future enhancements,
or not? Sorry, I'm not much into this legal stuff, it alocates too many
neurons, and my resources are scarse... :-)

Marcelo Jaccoud Amaral
Petrobrás (http://www.petrobras.com.br)
I'm not tense, I'm just terribly, terribly alert.

                                               Para:     [EMAIL PROTECTED]        
                      18/02/2003 19:23         cc:                                     
                      Favor responder a        Assunto:  Re: Licence issues in 
hyphenation patterns            

> I didn't modify the old pt.xml file, I wrote a new one entirely from
> scratch. ...

> The original TEX file was made by Paulo Rezende, now back in Brazil (at
> Unicamp), and it was converted by Paulo Soares (working presently in
> Portugal), who I contacted at the time. He posted no objection of us
> replacing the file.

Asking him was nice, but from a legal point of view, he can't sue
anyone for replacing his file by something else :-)


To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]

To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]

Reply via email to