h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

Since the recent upgrade, we've been experiencing problems with the search. In order to rectify this, the DNA team is currently re-indexing the search. This re-index of content across all DNA sites will take 2-3 days to complete.

During this time the search engine will work erratically. Only content that has been indexed will appear, so you may find that your searches doesn't retrieve everything you're looking for just now. Results will depend on whether the item you are searching for has been indexed yet or not.

During the re-index the servers are likely to be slower than usual. We apologise for this unavoidable reduction in performance.

Sorry, I'll compose myself.

I hope this all goes well.

Will it work quicker thereafter?

Not because of anything search is doing. When the server is indexing, it's doing a lot of disk accessing, so the database has to work harder to fetch its data, so all requests take a little longer. When the indexing finishes, it would go back to normal.

However, I've just been doing some work on the database, and I've improved a few slow-running queries, so you might actually notice a tiny improvement in speed. Difficult to say, though, since there are so many factors which can lead to slow pages, and this is just one of them.

sounds very good ! &

Some Frequently Questioned Answers:

<>

Not as such. However, it *will* be less slow, which is pretty much the same thing, except on thethird Sunday after a blue moon.

<>

s.....o.....u.....n.....d......s....
....l.....i......k.....e...
a.........l......o.......t
o......f.......
w.....o.....r......k.....

<< When the indexing finishes, it would go back to normal.>>

Oh, no! Not normal! Anything but that! I beg you to
reconsider!

<>

Tiny improvements in speed? Be still, my heart!

<< Difficult to say, though, since there are so many factors which can lead to slow pages, and this is just one of them.>>

I thought the slowness was caused by the same extraterrestrials that are causing our global warming and stealing half of our socks.

Top half or bottom half?

B

Front half.

We get a very surreal class of sock stealing extraterrestrial in our neck of the woods.

Very surreal if you're wearing them round your neck.

Or your wood, for that matter...

B

What bothers me most is the concept of wood having a neck in the first place.

Ia ia shub-niggurath and all that.

What I want to know is do the new indexing queries strip out punctuation using the same algorithm as the search box queries do? If you're rebuilding the index this would be a good time to check this, as in the past putting in an exact article title with punctuation got a very low hit %age, while leaving the punctuation out (even though it exists in the article title) got a much higher %age.

We have absolutely no idea what algorithm it uses - we're using Microsoft Index Server because it comes out of the box with SQL Server.

We're looking at how we can improve the search (either by optimising current search or using a different solution) right now, so in a month or so we might have some improvements to roll out.

Traveller in Time looking for clues
"Is there a list with not indexed words? "

Yes there is, and we use that, but the subtleties of how it breaks words aren't obvious. Having said that, our algorithms for stripping out invalid characters could definitely be improved (as I think Peet has pointed out before, prefixing a search with a double quote, but not closing the double quote, will not return any results).

Traveller in Time jumping another entry (A1923978) (But I think in monochrome the hooks show it is a link)
"If I dare to forget a double quote in I get an error by the parser?

No, serious, the 'Phrase Search' routine uses all kinds of brackets and quotes, I tried to figure out once.



"Monty Python" (only one python language in results)
Monty Python (pyhthon language in first page)

{Monty Python} second result
[Monty Python]
(Monty Python)
'Monty Python'

*Monty Python* jet another result not unlike second
!Monty Python!
@Monty Python@
%Monty Python%

"Monty Python is an error



I will have a local search if I still have the old statistics."

**unsubscribes due to technical bafflement**

B

*is too stubborn to unsubscribe*

*not sure anyone but Paul and Master B've been using English anyway*

Traveller in Time seeking cameras on 'Tower Bridge'
"Badgers everywhere < <./>/nature</.> > they announced some badgers in tonights show."

Verily it is written:

"System Update
A quick method of trashing ALL of your software."

and:
"For every action, there is an equal and opposite malfunction"

and:
"A complex system that does not work is invariably found to have evolved from a simpler system that worked just fine"

Traveller in Time on his head
"A functioning system is written as serveral milion bits either one or zero. There are only two sequences we can be absolutely sure of what it does. . 0 or 1.

The chances getting them in any order doing as much as this site is just a miracle."