One of the reasons that I "didn't mention" the notion of index threads is that an index thread about threads about accordions is just another thread about accordions, which will be found in a search for threads about accordions along with all the other threads about accordions, some of which may be in the index thread about threads about accordions and some of which may not be in the index thread.

And there's the additional question: "Should the Index Thread: Accordions be listed in the Index thread, since it's about accordions?"

This is such a fuzzy concept, one may easily see why I didn't mention it.

The idea of a search, or of a cross-linking index, that only returns selected results is somewhat like what Google does. Only the first few results, based on "most frequently accessed" and/or "most frequently linked to" are returned in the list of results. Google brags that 289,763,429 results were found, but will only show you the first hundred or so (it varies); and there is NO WAY that you can use Google in any effective way to get to any of the rest of them.

The Google method is fine for "surf-searchers" but is NOT ACCEPTABLE for any "academic search." It may be useful to find a few bits of "cream-amid-scum-off-the-top" for starting a ¹real search, but it's a bit "too yuppie" if you really need anything other than "pop" comments.

For some uses at Mudcat, a limited return list might be helpful; but I would suggest that due consideration be given to the need of some users to be able to find all of the items. The Googlish implementation prevents doing that.

¹ There are no generally accessible search engines for doing "real searches," that I've found, that aren't limited to extremely narrow interests, and/or that require "subscription" (with fees) to some database organization. Genealogy.com is one, and there are several "legal databases," with very high fees. Academic databases typically require you to be a registered student, and may have some areas limited to faculty. Getting a MIL-SPEC is nearly impossible unless you have a contract with the government. IEEE, SAE, AMA etc Documents and Standards are inaccessible unless you're a member, and even then may require substantial fees to see individual items.

I will agree with Max, subject to further consideration, that the long lists of "related articles" that appear on some threads are more intrusive than helpful. Putting them in a drop-down list, or simply inserting a "Click Here to find previous threads on XXX" might also work. If a link to a search (a Click Here) is used, it probably should be coded to open in a new window, so that newbies can return easily to the original page. If a simple link can access the list of associated threads, most here will know that they can right-click and choose whether to open in a new window, but persons handling "newness of the site" may not think to. (Does anyone not use a mouse when browsing?)

Drop-down boxes can be configured in several ways. With some that I encounter frequently, when you click on an item in the drop-down list, the selected link executes automatically. If this mode is chosen, the target should open in a new window (for this application). A disadvantage of this mode is that a "mis-click" takes you somewhere other than where you want.

The drop-down form most used here lets you make a selection, confirm that the right selection is in the window, and you click a "Go" button to execute the link. The Go button does not usually allow the option to "open in new window." In most cases where the link is to "additional information," if no choice is permitted, it would be my preference to open a new window. There are several places on the 'cat where "new window" would be inappropriate, since the choice is to a "destination," so some analysis of the use intended is needed when the function is set up.

It is my "understanding" (I may exaggerate by using the word) that one of the things that has delayed the appearance of new items in the DT is that the DT requires someone to "tag" items to have them appear in search results or to make Index Lists. The reference to "cross-links" for associating related threads indicates a similar Indexing rather than a simple text-search of the database. There are advantages and disadvantages to either method, and I don't have sufficient knowledge of the structures used by Mudcat and the DT to suggest changes; but it might be appropriate for some functions to "recreate" lists from text-search (like the Refresh with Filter?) rather than requiring someone to guess what needs to be "findable" and to create indexes/cross-links in advance.

John