On May 1, 2005, at 6:43 AM, Krzysztof Czaja wrote:
hi again,
Hans-Christoph Steiner wrote: ...
As for HTML docs, I don't understand why you want to separate the patches from the text. Yes, the text formatting abilities of Pd are
but you want them separate too... a reference page should be kept in a separate window anyway, so what is wrong about that window being controlled by another application?
If we can do it in Pd, why not do it in Pd? The example patches that go with the text would need to be opened from the browser (or whatever). There are a number of problems with launching .pd files from other applications, many of which are platform-specific and browser-specific. For example, on MacOSX, if you double-click a .pd file and Pd is not running, it loads the .pd file first, before loading the .pdrc or the prefs. So it'll the patch will throw errors if it needs anything from a non-standard path or a lib. This might affect other platforms too.
Plus, I think MacOSX is the only platform that automatically gets the file associations set up by Pd. On GNU/Linux and Windows, you need to manually create the file associations. That means for newbies, we'd have a lot of requests on the lists of how to do that. And each OS and each browser does it differently. Plus GNOME and KDE do it differently as well. If we write the references in Pd, we don't need to do anything to have the example patches launch in Pd.
However, my point is not about using a web browser, but about storing a reference page as a single parsable resource, instead of breaking it into several unparsable comments (same goes for any other type of longer self-contained text).
I am not sure what you mean by "unparsable comments". Using a multi-line regexp, you can grab all chunks that start with '#X text [0-9-]* [0-9-]*' and ends with a non-escaped ';'. The '#X text [0-9-]* [0-9-]*' would be equivalent to an HTML <p> and ';' to a </p> Start that comment with a keyword, they its quite easy to assign that text to that keyword in a hash or whatever. The regexps could be used in Tcl to be cross-platform. I don't think it'd be any harder to parse a .pd file than it would be to parse a .html file for our uses.
.hc
The way it is displayed is quite another matter. I am proposing the external browser, as the easiest, and the most old-fashioned-Pdish way to do it. Another option is including some kind of a browser widget in pd-gui, which you seem proposing:
...
of that. With a little more work on the [comment] object, I think we could have all of the text formatting capabilities that we'd really need to make very nice, linked help patches with texts, all within Pd.
Krzysztof
________________________________________________________________________ ____