Setting up a new Wiki
HOW NOT TO START A PROJECT
I rashly bought a Swiss domain jaypri.ch when I really don't need personalised email and already have a free static homepage that I don't use. Now it's a commitment! I should do something with this ip address but what?
WHAT IS A GOOD QUESTION
I've been messing around recently with G'MIC which is an easily script-able plugin tool for GIMP. Mostly the work of David Tschumperlé it has attracted a creative user-base (meet here) and is a complete enough tool to prototype new image processing algorithms with.
So here goes, let's create a Wiki and learn something while trying.
STEP 1. Who's my audience?
So taking the GMIC example; it already has clear linear documentation for the command line version that explains how to script for the GIMP plug-in and many routines have useful notes that appear to a user.
So I'm imagining a typical user of this Wiki will come wanting to perform a particular ImageProcessingTask and not necessarily to be asking about syntax.
STEP 2. How will they find what they want?
Well I'll have to put myself back in the shoes of someone coming across this for the first time *.
The main thing a good Wiki can do is increase the likelihood of finding related things of interest. The methods supported will be:
- I hope a search function is included or will be easy to implement
I'd like also to see its dual, a log of what terms were actually used in a search to find a page if possible
- Similarly a sortable glossary would be nice
including aliases and keyword tagging each page will be an important admin task I suspect, I think it will be possible to use that meta-data in aggregate to build and filter a relevant list, a good way to find a topic when you don't know exactly what it's technical term is
- I want to enforce certain pages into categories, I know a wiki supports that and "toc"
to seed the site then I plan to upload versions of the existing documentation and filters according to the GIMP plugin's own menu .. as I want to make it easy for people to go from the Wiki and apply what they've learned
- If the Wiki grows then probably it will need sections of "type" with their own recognisable templates
for instance a single script command or a filter can be differentiated from a concept or ways to achieve an effect.
Step 3 Stickiness, engagement.
I have no idea if this will be well received by the community. I'm hosting the site at my own expense for six months, no advertising. If it proves so popular it becomes expensive then I'll have to offer to covert it to a static site which is cheaper or free.
Anyway I think it's worth a try to curate the rich history of discussions already in flickr and gimpchat, the extensive size of it's library of routines makes it easy to miss things.
* HOW I FOUND GMIC
Personally I'm fond of the usual "search and hope" method a Google habit encourages of finding solutions to a problem and I came across GMIC for the first time while trying to suppress the nasty digital grain and noise in an airborne photo someone took with their iPhone.
It turned out this plugin could do "anisotropic blurring" which is what I thought I wanted, but another filter "Enhancement/Smooth [thin brush]" was even better. Anyway I loaded the plugin and once I'd made the new Desktop Wallpaper for myself forgot all about it - didn't even consider looking at how to write its script.
It was not until I was reading up on artificial intelligence and deep belief networks - totally unconnected you might think - and via a comparison of BlockMatching3d de-noise algorithm a trained network found also a lecture in English by @ronounours in April 2009 explaining Patch-based smoothing http://videos.rennes.inria.fr/seminaire_Irisa/Vista/
This sparked interest to look at the source code for CImg and to compile GMIC (not easy with my limited understanding of cygwin/mingw and x86_64). Anyway the script is fun and command-line is very easy so I've not looked back since!
I rashly bought a Swiss domain jaypri.ch when I really don't need personalised email and already have a free static homepage that I don't use. Now it's a commitment! I should do something with this ip address but what?
WHAT IS A GOOD QUESTION
I've been messing around recently with G'MIC which is an easily script-able plugin tool for GIMP. Mostly the work of David Tschumperlé it has attracted a creative user-base (meet here) and is a complete enough tool to prototype new image processing algorithms with.
So here goes, let's create a Wiki and learn something while trying.
STEP 1. Who's my audience?
So taking the GMIC example; it already has clear linear documentation for the command line version that explains how to script for the GIMP plug-in and many routines have useful notes that appear to a user.
So I'm imagining a typical user of this Wiki will come wanting to perform a particular ImageProcessingTask and not necessarily to be asking about syntax.
STEP 2. How will they find what they want?
Well I'll have to put myself back in the shoes of someone coming across this for the first time *.
The main thing a good Wiki can do is increase the likelihood of finding related things of interest. The methods supported will be:
- I hope a search function is included or will be easy to implement
I'd like also to see its dual, a log of what terms were actually used in a search to find a page if possible
- Similarly a sortable glossary would be nice
including aliases and keyword tagging each page will be an important admin task I suspect, I think it will be possible to use that meta-data in aggregate to build and filter a relevant list, a good way to find a topic when you don't know exactly what it's technical term is
- I want to enforce certain pages into categories, I know a wiki supports that and "toc"
to seed the site then I plan to upload versions of the existing documentation and filters according to the GIMP plugin's own menu .. as I want to make it easy for people to go from the Wiki and apply what they've learned
- If the Wiki grows then probably it will need sections of "type" with their own recognisable templates
for instance a single script command or a filter can be differentiated from a concept or ways to achieve an effect.
Step 3 Stickiness, engagement.
I have no idea if this will be well received by the community. I'm hosting the site at my own expense for six months, no advertising. If it proves so popular it becomes expensive then I'll have to offer to covert it to a static site which is cheaper or free.
Anyway I think it's worth a try to curate the rich history of discussions already in flickr and gimpchat, the extensive size of it's library of routines makes it easy to miss things.
* HOW I FOUND GMIC
Personally I'm fond of the usual "search and hope" method a Google habit encourages of finding solutions to a problem and I came across GMIC for the first time while trying to suppress the nasty digital grain and noise in an airborne photo someone took with their iPhone.
It turned out this plugin could do "anisotropic blurring" which is what I thought I wanted, but another filter "Enhancement/Smooth [thin brush]" was even better. Anyway I loaded the plugin and once I'd made the new Desktop Wallpaper for myself forgot all about it - didn't even consider looking at how to write its script.
It was not until I was reading up on artificial intelligence and deep belief networks - totally unconnected you might think - and via a comparison of BlockMatching3d de-noise algorithm a trained network found also a lecture in English by @ronounours in April 2009 explaining Patch-based smoothing http://videos.rennes.inria.fr/seminaire_Irisa/Vista/
This sparked interest to look at the source code for CImg and to compile GMIC (not easy with my limited understanding of cygwin/mingw and x86_64). Anyway the script is fun and command-line is very easy so I've not looked back since!
Comments
Post a Comment