Consider that we give topics only to nouns (to answer the question "What") and have discussions as verbs relating to those nouns. It would require a SubWeb for each discussion under the NounTopic and some funky modification to the WikiSyntax so that Nouns#TopicsOnlyFor could be written as TopicsOnlyFor##Nouns.

Hence you can go find all topics about the noun Noun.

-- MartinCleaver - 19 Jun 2001

Also of interest: http://www.wikipedia.com/wiki/Naming_conventions (or is that Wikipedia:Naming_conventions ?)

-- MartinCleaver - 26 Jul 2001

More on this: by using the identities of the most pertinant nouns as the identity (i.e. WikiWord) for topics that discuss where those nouns are considered in conjunction, that combination of nouns has to end up being discussed in the same place. This is excellent for ensuring convergence of content.

I am imagining a situation where you separate the nouns with commas, such as:

• TWiki,WindowsPlatform

And to avoid duplication use a rule such as always sorting the nouns alphabetically in the WikiWord.

For syntactic sugar add that the display of the topic can be different to the underlying WikiWord (i.e. we can display TWiki,WindowsPlatform as TWikiOnWindows) that and while you are there you might as well have different displays for different languages (e.g. SpanishDisplay: TWikiEnMSVentana).

-- MartinCleaver - 28 Aug 2001

Furthermore, by having the topic titles named predictably we lessen the chance of slight but otherwise inevitable variations such as QuickPlaceVsTWiki, TWikiVsQuickPlace, TWikiAndQuickplace, TWikiComparedToQuickplace, TWikiMeetsQuickplace. These all get to be discussed at the topic Quickplace,TWiki.

-- MartinCleaver - 29 Aug 2001

Ahha, I'm starting to see what you're getting at! My comment is a little off topic, but this is the reason I'd like to see a better search engine for TWiki. If you could search for QuickPlace "near" TWiki (or "w/5" words), you could find the appropriate topic(s) quickly no matter which way they were named.

I prefer to search for what I want in the manner described above than try to remember (or reconstruct) a topic name. How about other people that look for information in TWiki -- would a search engine with more capability or a more systematic approach to topic naming be more useful?

I'm not trying to start any kind of flame war -- just trying to see what others think. I see nothing wrong with a more systematic method of naming topics, especially since it can be done immediately with no additional development effort (unless we want the systematic method to be enforced by the software somehow). I just know that no matter how well though out the system is, I will not guess the appropriate topic, and the search strings that I enter into the TWiki search engine will produce excessively long lists of hits and, at minimum, I will have to read through that list of topic names.

(Just trying to increase the enthusiasm for a better search engine, which I haven't been able find and install myself. (Nor have I got TWiki (WikiLearn) on SourceForge fully working yet.))

-- RandyKramer - 29 Aug 2001

I certainly agree that a better search engine is useful. The main approach we've used at my company has been to use HtDig. I think it would be good see some plugin hooks in TWiki so that the searching can be changed in different installation. This would include hooks when updates are done so an indexer can be put in place. I can't see how we can get people to stick to a specific naming convention, I think searching is the way to go.

-- JohnTalintyre - 29 Aug 2001

An example of where separating nouns by commas would be useful is the existance of two topics on TWikiDotOrg - MainframeTWiki and TWikiOnMainframe. The key ideas in these two are TWiki and Mainframe. The possibility of the existance of both topics is due to the fact that you can write in either order the combination of the WikiWords that represent these ideas. And, the more nouns that occur the more likely that conversation about the intersection of these ideas will get represented in multiple topics.

So, the proposition that I made on 28 August is that topic identifiers (i.e. WikiWords) that represent the convergence of these ideas be comprised by compounding the WikiWords of the underlying ideas in a predictable way, i.e. alphabetically ordered and separated by commas.

A further requirement of this idea is that of SingletonWikiWords, because sooner or later you (may) need to define WikiWords down to their single consituent parts.

And, just to throw a spanner in the works for UNIX TWiki users, I believe that TopicCaseInsensitivity is necessary, if not least because:

1. People are especially inconsistant about how they are forced to WikiWord an otherwise single word. (e.g. Ontology as OnTology or OntoLogy)
2. Some words are written inconsistantly in natural language, e.g. MicroSoft vs. Microsoft

-- MartinCleaver - 19 Nov 2001