Have a dig around CPAN. There are modules in the Lingua name space that may help (although I suspect you are a step or two beyond what they are aimed at). There are a lot of modules aimed at facilitating WordNet access however and a small number for VerbNet.
Perl is environmentally friendly - it saves trees
| [reply] |
| [reply] |
framboise:
Why are you tempted to rewrite it in perl? Don't get me wrong--I *love* perl. But I hate to rewrite anything when I have it already written.
I don't remember much PHP (it's been about 7 years, so anything I remembered would be out of date anyway), so I don't know what benefits you'd gain. But all languages have a few potholes here and there. So when you rewrite it, unless you're going to rearchitect bits of it as you go, you'll find out that you'll have to dodge a few perl potholes, and your code would be littered with the remains of translated PHP potholes. That (and the time investment) would probably offset any gains perl might give you.
Just askin'.....
...roboticus | [reply] |
Most of the work that I've put into the project has been research and databasing, so the PHP I've written isn't so extensive that it would be incredibly hard to rewrite, and most of it will need revision as I try to incorporate it into a larger system.
Perl is more dynamic than the PHP that's locked up in my server and from what I've read can probably be integrated more easily into other systems. This might actually become more of an issue later if I work with systemic grammars. Perl modules will allow me to cut some corners, and I'm all for that. I already see that there are some modules that do almost exactly what my stuff does. Lastly, although it at first didn't seem like much of a factor, the Perl community is more centralized and geared towards this type of work. I found very little documentation on NPL+PHP, because that's just not what PHP is about.
| [reply] |
Not only are there some strong NLP people in the Perl community, but the language was designed by a man educated as a linguist. The initial role of Perl was text processing. It has grown well beyond just that, but it still handles text better, faster, and more comfortably for the programmer than many other languages.
I'm not personally very well versed with the linguistics modules on CPAN, but I've heard very good things about some of them. I've looked briefly at several because they fascinate me (I intended to be a linguist myself at one point), and I can tell you there's much more value on CPAN for NLP from what I can tell than there is out there for PHP. In the interests of full disclosure, Python and Java also seem to be well-represented in the NLP arena, though.
As long as you're looking at alternatives, I'd suggest looking at http://opennlp.sourceforge.net/projects.html, where a number of Open-Source NLP packages are described.
| [reply] |