|
From: | Juergen Sauermann |
Subject: | Re: [Bug-apl] Questions about GNU APL |
Date: | Tue, 01 Apr 2014 12:30:16 +0200 |
User-agent: | Mozilla/5.0 (X11; Linux i686; rv:17.0) Gecko/20130330 Thunderbird/17.0.5 |
Hi Blake,
i did not mean to say that every APL programmer should write her own file system interface, or that component file systems are not useful. What I was after is the way such a file system is integrated with GNU APL. My current model is that of C/C++. There you have the core language, say C, that is comparable to GNU APL. Then you have libraries for everything else including different databases, graphical stuff, math, and the like. These days the libraries together are far bigger than the languages, and it is the libraries more than the language that makes C/C++ usable everywhere. If we would use ⎕xx functions for integrating a component darabase (for example) then the API (i.e. the ⎕xx) function would become part of the GNU APL core language. That is not wise for a number of reasons: documentation, maintenance responsibility, modularity to name a few. It also makes programs using them less portable (even the encoding of the character ⎕ is different on different APL systems). The reason why I mentioned GNU APL native functions is that they are the intended mechanism to extend GNU APL (or in the above context) to build libraries that can be used (i.e. called) from GNU APL. And, unlike ⎕xx functions, they can be replaced by wrapper functions written in APL to port APL programs to other interpreters without the need to change every call of ⎕xx. The emacs mode written by Elias is an example of such a "library". It allows us to please emacs users without any (well - almost) impact on GNU APL itself. I believe a database interface should be done in the same fashion. So if you would like to provide a component database for GNU APL, please feel invited to do so. If there are functions missing in GNU APL then I will try to provide them. That was kind of the work-split between Elias and myself in the case of emacs mode. Best Regards, Jürgen Sauermann On 03/31/2014 05:00 PM, Blake McBride wrote:
|
[Prev in Thread] | Current Thread | [Next in Thread] |