Indexing - Optional?

David McAllan said:

I have read a lot of complaints about indexing and understand them all. However, I can see the merits of indexing and think it should be a must for L4.  We need to remember that this is a product for the future and hard disk space, while it may be a dilemma now is certainly not an issue on recently built machines.  Unfortunately, Logos is "caught between a rock and a hard place".

I don't know what it involves, but can I suggest that their is an OPTION to ignore/turn off indexing so that searching could be carried out as per L3.  In this way, those who have slow machines could search immediately and perhaps save some HD space, but for those who have both may be happy to persevere with the indexer.  Having the option seems to me to be the way forward.

David

I think it would be nice, however I am skeptical about the possibility, since it seams to me that Logos4 is using the index not only in searches, but in many other places that show simultaneously. I am anxious to hear from developers on that. (For me, I would use index even on the slower machine, just because it is slower). The only thing which would be good to have would be an option to postpone indexing to the more suitable time.

Bob Pritchett said:

The whole product is designed around a one-index library that we can search quickly. If you prefer the old 3.0 style of searcing each book separately -- sometimes for 30 minutes or more, to cover the whole library -- then the best solution is to use 3.0 on that machine.

-- Bob

I understand the reason for indexing, but I do question the need to re-index on each update.  It seems that the index should be capable of being imported into the next release.  I tend to think that the only need to re-index would be when there is a change in resources (adding or deleting a resource).  This would, of course, require indexing on demand.

George Somsel said:

I understand the reason for indexing, but I do question the need to re-index on each update.  It seems that the index should be capable of being imported into the next release.  I tend to think that the only need to re-index would be when there is a change in resources (adding or deleting a resource).  This would, of course, require indexing on demand.

Indeed. But why a global re-index? Surely Google et al. don't re-index the entire Internet every time a new website appears or a blog post is written.

Oh to have 20+ gigs of ram and keep the whole of Libronix resources in RAM the way Google keeps the internet in RAM.  (Seriously, that's a sick amount of RAM!)