erwanm / tw-aggregator

a system to automatically aggregate TiddlyWiki content from a collection of public wikis
6 stars 2 forks source link

Categorizing the tags, channels and/or collections, customized aggregated content #77

Open twMat opened 9 years ago

twMat commented 9 years ago

The idea to permalink specific tags is brilliant - thanks @felixhayashi - and should prove a definite carrot to use the tw-aggregator.

However, the tags list is very long, and growing. Including a bunch of "weird ones". And, as indicated in the instructions to Bookmark tags, someone interested in images can find tag "Image" and "image" to bookmark. But it is very reasonable that he is also interested in tags; "img", "gallery", "background", "graphics", "picture", "pic" ...and so on. And then double the list by including capital initial letter. Etc.

Usability on this would greatly increase if the tags were somehow categorized - eg. the above variants could all go into one category so that only this has to be bookmarked.

...buuut, it may be that the tag "graphic content" is not something you'd want to include. So there is need for different categories containing overlapping subsets of tags.

Now, how could this be done without putting a workload on the aggregator host (monsieur @erwanm in this case)?

Well, this is a job for "the people"! Members can create channels with what tags are interesting and others can bookmark/reuse them. A bit like on youtube I think.

The question is how such a "channel" is available for viisitors to the aggregator. IMO it seems like a perfect thing to couple with your author profile tiddler.

Whaddyathink?

erwanm commented 9 years ago

That's again a really interesting idea @twMat, and it made me think of something more general. To explain it here are the steps my little brain went through:

  1. let's assume wiki authors can have "channel tiddlers" which contain several tags like you said, so that the aggregator can collect these channels and build the permaview link itself.
  2. but why only build the link? the aggregator can simply create a single tiddler which will display the list of all tiddlers corresponding to any of the tags, making it even easier for users.
  3. Then why not extend the filter even more? Instead of only the union of several tags, a channel could be any filter based on tags, author, wiki, and could even contain a search operator for a specific term, etc.

This way users would have access to fully customizable lists of community tiddlers, in a single tiddler that they can bookmark, so that all that it requires is a single click. Nice, right? And it's actually quite easy to implement, since TW would do most of the work.

twMat commented 9 years ago

Add: since writing the following I've realized the following may not be quite doable. It is based on the assumption that what a user gets when he downloads a new TW via tw.com is assembled from individual tiddlers. I think this is how the tw.com site is made, but this doesn't have to be the case for the download file. I will ask about this and return on the matter.


Hehe... and not only that!

This would maybe let people set up customized editions/applications - i.e FULL TW's could be aggregated, but specialized according to some particular need/preference/application. I.e some kind of link-list for downloading stuff....

...right?

Take a peek at @Drakor 's Customizer for an idea for UI.

Users could choose e.g

... or even...

... At a later stage (to not make this overwhelming... I have a tendency to let my fantasy run) possibly this may be interesting:

In a recent thread on the boards, apropos @pmario mentioning that he finds the core to be bloated, I asked what it would take to remove/externalize core features. For instance, if I don't want the New Journal and New Journal Here buttons/features. @Jermolene expalined that the best thing was probably to overwrite then with blank tiddlers. The ideal, is of course(?), to decide already on download if one wants them or not. In the (specific and arbitrary) case of "New Journal", this woudl be simple but it might be trickier for other components that are more critical for overall functionality. I think it is an area worth exploring, later on.

BTW, maybe the ideas on recipes and bags from TiddlyWeb are of relevance here? I know basically nothing about TiddlyWeb, but I'm thinking that just maybe "they" have thought through this idea of concocting tiddlers int o a TW in some clever ways that are worth considering.

erwanm commented 9 years ago

Wow, you are always thinking further! ;)

That's interesting, but I doubt I would be able to go this far technically myself.

However I will keep in mind the idea of generating a customized TW with aggregated content: that might be a way to solve the performance problem with the content growing, since users could select the content they are interested in with filters (e.g. only from authors X and Y, only the N latest tiddlers, etc.). if they prefer to have a fast search they would select less content, for instance.

erwanm commented 9 years ago

Tobias' views about collections/channels: https://groups.google.com/d/msg/Tiddlywiki/nluco-cZ9-s/8ECSaRlEAgAJ

Jermolene commented 9 years ago

The ideal, is of course(?), to decide already on download if one wants them or not. In the (specific and arbitrary) case of "New Journal", this woudl be simple but it might be trickier for other components that are more critical for overall functionality. I think it is an area worth exploring, later on.

That makes it sound as if a custom core plugin is assembled for each downloaded wiki. If so, I really don't think that that is a good idea - see the discussion elsewhere about custom versions of the core plugin.

The way to customise a TiddlyWiki during the download process is by adding things, not taking things away.

erwanm commented 9 years ago

I'll keep the idea of generating custom wikis on hold for the moment, but I hope to implement custom filters soon-ish.