[PD] multi-language help patches

[Lucas Cordiviola]

Can it be done with puredata.info?

Mensaje telepatico asistido por maquinas.

From: danomatika at gmail.com
Date: Fri, 26 Feb 2016 17:35:58 -0700
To: jancsika at yahoo.com
CC: pd-list at lists.iem.at
Subject: Re: [PD] multi-language help patches

I'd look at how the processing team has put their site/reference together:https://github.com/processing/processing-docs
It's basically using a php engine with content written and parsed in xml. Each function/class to be documented has it's own file and each file starts form a template which the author then just needs to fill out. I'm not suggesting using xml, etc but I like the layout from an editing & parsing perspective.
And this is the processing reference, which I feel is a great model for a teaching/arts environment (aka good balance towards readability for beginners and the important info for pros). This section is also included when you download Processing and opened automatically via a link form the Processing IDE itself as well as contextual menu when hilghighting functions & classes in the editor:
https://processing.org/reference/
Openframeworks has been going in that direction, but the codebase is a lot larger and the underlying C++ requires more details to explain, etc hence the recent DocSprint. So far, the examples have been refreshed and lots of useful snippets & longer form tutorials/books have been put together:
New Learning page from last weekend: http://openframeworks.cc/learning/Existing reference: http://openframeworks.cc/documentation/
Here's the current OF website using markdown parsing & generation:https://github.com/openframeworks/ofSite
And here's a new approach to generating the OF documentation as a work in progress:https://github.com/halfdanj/ofdocumentationgenerator
Here's a couple docs and workflows that have come out of and/or been generated by the DocSprint work:
https://hackpad.com/API-Documentation-Overview-wCQ4J5hrBbW
https://github.com/openframeworks/openFrameworks/wiki/Examples-Contribution-Process-Flow
I feel one of the best aspects of PD are the examples via help patches so maybe splitting things up outside of PD might work against that? However, this would allow for a place for larger scale information as well as images, embed, etc.

--------
Dan Wilcox
@danomatika
danomatika.com
robotcowboy.com



On Feb 26, 2016, at 5:20 PM, Jonathan Wilkes <jancsika at yahoo.com> wrote:Do you have a spec from that sprint?

    On Friday, February 26, 2016 7:08 PM, Dan Wilcox <danomatika at gmail.com> wrote:
  

 Also, my thinking is going in this direction as we’re dealing with the same issues in the OpenFrameworks community. My uni department just hosted an OF DocSprint last weekend and we spent a good amount of time wrangling how best to integrate a Markdown + Doxygen generated reference system.Of course pure data patch files and C++ source files are somewhat different, but I feel there are the same issues to solve such as what requires the most maintenance, works on all platforms, and is easy for non developer contributors to use. It’s one thing to build a custom system (we did) and quite another to get people to pitch in and fill the content in. I just wouldn’t want anyone to spend a lot of time making something admittedly cool and built into the canvas but, in the end, may not be leveraged by the community the same way a portable, easy to edit, cross platform standard might.
--------Dan Wilcox at danomatikadanomatika.comrobotcowboy.com


On Feb 26, 2016, at 5:01 PM, Dan Wilcox <danomatika at gmail.com> wrote:Ok, so which html reference system should I leverage here?Probably something using css and an html template that make it easy for people to fill out. I’d say 1 main html file for each object to document w/ room for sub pages if needed. Different languages can live in different folders.The nice thing about this approach is lots of people can edit html, there are plenty of designers, the files can be rendered by pretty much anything, etc. Another option is to have a templating system that uses Markdown, etc and just renders to html. It can then live in it’s own source repository for shared work and be used as a basis for online as well as distributed documentation.Maybe a good start would be to look at the pure data object database/wiki that is around somewhere. I can’t find the link off the top of my head.Where will the html files get stored, and how do we get from clicking the link in the help patch (I'm assuming we're still using the current help patches to show a simple demo of the object) to opening the html doc in the correct language?Just like opening a help patch with a context menu option or maybe links we can open from the patch itself. Use the current help paths for searching and use tcl to launch the path in the system web browser if found.I’d say the most useful thing would be add linking between patches and external files (html, etc) in general. I believe Hans had this in extended for the pd-doc stuff.I’m suggesting this approach partially so you/we don’t end up reinventing the wheel. A custom, integrated system would be *nice* but I feel that will require too much backend work to build and them probably too much work to maintain/extend in the future. HTML+CSS has the option of being loaded into a web view within TK I imagine, so another option would be a side pane or extra window that can open up right in PD. I’d suggest staying away from building extra widgets etc to render a custom approach within the patch itself.
--------Dan Wilcox at danomatikadanomatika.comrobotcowboy.com


On Feb 26, 2016, at 4:44 PM, Jonathan Wilkes <jancsika at yahoo.com> wrote:-Jonathan     On Friday, February 26, 2016 4:34 PM, Dan Wilcox <danomatika at gmail.com> wrote:   I think what implying is that maybe Pd *doesn’t* need to handle it. Simply, Pd could open a local webpage, similar to how the Processing “Find in reference” context menu option works when highlighting a function in the editor.Not to say you/we can’t work out a file format/system to handle alot of this, but I’m thinking that html reference already works well for many other contexts an doesn’t require building new formats/systems to solve alot of the same problems.
--------Dan Wilcox at danomatikadanomatika.comrobotcowboy.com


On Feb 26, 2016, at 2:08 PM, Jonathan Wilkes <jancsika at yahoo.com> wrote:html could be leveraged, but I'm really looking for a spec for how Pd handles it.  Is it a GUI widget?  An abstraction?  A canvas method?  A new "#" directive?Do the translations get saved along with the help patch, or are they stored in a directory and fetched when needed?  Etc.-Jonathan     On Friday, February 26, 2016 1:02 PM, Dan Wilcox <danomatika at gmail.com> wrote:   I'll implement any *clear* spec for multi-language help patches someone comes up with with the following constraints:1. it separates design from content.2. in only requires documentation writers to care about content.3. it does not pigeonhole help patches into having a single, ugly design4. documentation writers will be guaranteed that whatever they write, it won't overlap patch content.5. it is maintainable and scalableSounds like .html.
--------Dan Wilcox at danomatikadanomatika.comrobotcowboy.com_______________________________________________Pd-list@lists.iem.at mailing listUNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list          

     

_______________________________________________
Pd-list at lists.iem.at mailing list
UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20160227/6e685a47/attachment-0001.html>