TWC include plugin

[stefan infp]

Using TWC with the include plugin is what it is necessary to create huge interlinked wikis. The question is as the wiki collection grows, also the main index wiki grows and becomes heavy loaded disappearing the benefits of using of the include plugin isn't it? Thousands of wiki files indexed in a main wiki can consume the system memory and freeze the browser, I suppose. How are the indexed wikis loaded? As I see, the tags are loaded all in a tag list but are the linked tiddlers loaded at the opening of the main index wiki or are they loaded each one when opened after clicked on them in the search results?

[Tobias Beer]

 Hi Stefan,

 

Thousands of wiki files indexed in a main wiki can consume the system memory and freeze the browser, I suppose.

Surely, there are limits to any approach. For thousands of wikis I would not expect any browser based approach to work that aggregation well. This would call for an entirely different stack of interlinked software processes, both on the model end (database) and the controller (watching for changes to individual wikis and triggering / pushing updates to the global aggregation db).

So, if you want to manage "thousands of wikis" that way, I'd argue neither TW2 nor TW5 cater for your needs in their browser based version.

Best wishes,

— tb

[stefan infp]

Hi Tobias, 

Thank you for the quick answer. So if I want to work with TW I should use the node version or multiple small TW5 wikis and index and search their content with Windows Explorer. Node is not so user friendly and I would need a manual to use TW in command line under Node (also it stores all embed media files in a single folder which could lead with a huge mess). Does TW5 on node.js can face the same problem with huge amount of text and image tiddlers? Thanks.

[Tobias Beer]

Hi again, Stefan,

 

Thank you for the quick answer. So if I want to work with TW I should use the node version or multiple small TW5 wikis and index and search their content with Windows Explorer.

 

The benefit of using node would be to have neatly chunked tid files, rather than wikis to search. What is missing as of yet is a parser / indexer that semantically searches tid files, e.g. certain fields, like tags or title or text, or date fields, etc..  A CLI (and possibly a ui on top) would surely majorly enhance the node experience.

Node is not so user friendly and I would need a manual to use TW in command line under Node.

I certainly see how using a CLI comes with a barrier and requires a greater willingness to learn and remember and perhaps a need to often use certain commands, to not having to constantly look up the right one and its spec.

 (also it stores all embed media files in a single folder which could lead with a huge mess).

There may be a default storage location for embedded content. However, you are not bound to use that, you are quite free to organize both tiddlers and external files in any way you find suitable to your project(s). In fact, your folder structure and the eventual titles / addresses of those tiddler sin your wiki are entirely independent.

Does TW5 on node.js can face the same problem with huge amount of text and image tiddlers?

I think on node it is much easier to manage tiddlers in whichever wiki / folder structure you want and to cross include different components into whichever build / bundle / aggregated wiki you want to eventually create.

So, the complexity of your considerations sure calls for node. Perhaps usability comes with actually using it day to day. Also note that using the node version does not bar you from working in a browser environment. It merely adds the optional steps to organize your tiddlers in wiki folders on the (server-side) filesystem.

Best wishes,

— tb