Tech: Google link help needed!

Subject: Tech: Google link help needed!
From: GUEST,Alan Ross
Date: 05 Aug 06 - 11:43 AM

Could any of you wizards out there restore the link to infomation I deposited on the history of 2 of my late father Stewart Ross's Scottish Songs 'Here's to Scottish Whisky' aka 'Here's to Scottish Whiskey' (sic) and 'My Bonnie Maureen' (where I transribed the words). Very oddly when I type in the titles of the songs into Google there is no link to Mudcat cafe's stored pages, but if I pick out odd phrases from the pages it links them.

I'd be very grateful if somebody could rectify the situation and link the titles for research purposes. One Internet music encyclopedia, has my father's songs as been written by Ross, Stewart - Instead of his name being Stewart Ross - that's why I would like the correct details listed elsewhere in case confusion arises. At one stage google listed several hundred references to both songs and it was increasing - then suddenly they seemed to have been narrowed right down on an overnight basis!


Alan Ross

Subject: RE: Tech: Google link help needed!
From: Stilly River Sage
Date: 05 Aug 06 - 11:50 AM

That may be a Google issue and not a Mudcat issue. I hope you have good luck in someone helping you with that, but there has been a complaint about what Google (and other search engines) can or try to find within places like Mudcat.

SRS

Subject: RE: Tech: Google link help needed!
From: GUEST,Alan Ross
Date: 05 Aug 06 - 12:03 PM

Thanks SRS the link worked for ages - then overnight seemingly disappeared!   It's crazy that Google's engine is picking out phrases from Internet pages - yet the very title of the songs that the pages are about has ceased to be linked - so nobody would be able to find the history or lyrics to the material via the song title!
Alan

Subject: RE: Tech: Google link help needed!
From: JohnInKansas
Date: 05 Aug 06 - 01:25 PM

Google is a search engine, not a permanent index.

The actual details of how Google decides to return a particular web page in its results is a closely guarded secret, but a few things are pretty well accepted.

First, Google decides whether to create an index of a site based mostly on how many other sites link to it. Although the number of pages for which indexes are gathered is huge, it's still only a small fraction of the pages on the internet.

Second, Google does not necessarily make a "complete" index of every page. They use an arcane formula (or several) of their own design to determine what they can later "find" on a page that is indexed. The decision probably is based largely on how "popular" a scrap of information is likely to be.

Third, Google apparently sorts the "queries" by popularity. If lots of people are looking for something, the most commonly used "search terms" get priority. This sometimes means that a term less often searched for gets moved out of the system.

Fourth, Google sorts the "hits" by popularity. Once a term has been searched, the position of a particular "hit" in the list of results depends on how many people have clicked on that result. More clicks leads to appearance higher in the list. It's "impressive," I suppose, that Google always reports "100,000,000 results found" for a search; but it will only show you the first 200 or so and there is no way (that I've found) to see the other 99,999,800 results.

When Google has indexed a page, and your search finds a particular result, you cannot rely on always getting the same result in future searches. Google will update its scan of "popular" pages, and will change its indexes for those pages based on what people have clicked in previous results. A song title that few people "clicked" in earlier results can easily get dropped off the index, in favor of something else that more people "expressed an interest" in.

A fifth factor is that a web site that uses "extreme" measures to artificially increase its odds of getting listed in Google may get "black listed" by Google, and all results on the site's pages may be blocked from being displayed in search results. (Most likely this is done by moving its stuff down into the 99,999,800 results one can never see?)

If you come to mudcat, and use the searches on this site, your results should be "permanent" and a search that works once should always work. Mudcat can do something about a song that disappears from its own search machine - probably. But nobody here has any control over what Google reports, and Google's own system is designed to be constantly changing, and to report only the "most popular results."

John

Subject: RE: Tech: Google link help needed!
From: GUEST,Jon
Date: 05 Aug 06 - 01:40 PM

It's a matter of luck whether a Mudcat thread will get indexed on a Google pass - it depends what is visible when the search engine does it's job.

Put the link to the Mudcat thread on a permanant page somewhere on the Internet that Google can index and it should then pick up the Mudcat link again and keep it (at least if the other page really is permanent).

Subject: RE: Tech: Google link help needed!
From: Joe Offer
Date: 05 Aug 06 - 02:03 PM

Stewart Ross" whisky maureen brings up some interesting stuff - AND a Mudcat thread. Also, take a look at what comes up with "scottish whisky" "stewart ross" (but not with whiskey). Some combinations of phrases from that thread don't bring up anything on Google - which goes to show that Google is not all-encompassing.
Unfortunately, our search engine index hasn't been updated for a year, so some things don't turn up on our own forum search. Best to find threads using the Filter and searching for the thread title.
-Joe Offer-

Subject: RE: Tech: Google link help needed!
From: JohnInKansas
Date: 05 Aug 06 - 02:29 PM

For those who'd like a first hand view of how "search engines" work, it is suggested that one can get "close up and personal" via the Google Desktop Search utility, or the competing Microsoft Desktop Search.

Early versions of Google Desktop Search raised some security questions, and with my immaculately crafted file structure I didn't feel it was needed so I've never tried it on my machine. (The security concerns reportedly have been addressed.)

Having progressed to where a full Windows Explorer search of "My Computer" for a file containing a single word takes a bit over 2 hours to run, I was tempted when I came across a glowing recommendation for the Microsoft Desktop Search, so I did download it (about 4 MB, 20 minutes on my dial-up connection).

As with Google, this utility works by going through all the stuff you want to be able to search, and making an "index" of what's on your machine. Once the index is made, only the index gets searched when you look for something. (Supposedly, the index is continually updated in background.) The default setting is to index your "My Documents" folder and your email. Since I don't keep anything in "My Documents," I changed the "preferences" to index my hard drives, excluding system files, and to index my email. (2 hard drives, 149.5 GB total used of 280 GB total available, 530 MB email)

It took just over four days to complete building its index.

Apparently it applied "indexing rules" somewhat like Google's, since trial searches were very fast, but did not find known files using my usual searches on known file content. (i.e., the indexes were far from complete).

On completion of its "index build" it installed itself in place of Windows Explorer Search, making my previous "full text search" for anything it didn't index impossible. (Once it's installed and indexed, you can only search the index.)

The interface was dazzling, with previews of everything (that it found) and it was remarkably fast; but since it disabled more thorough search …

It took 4.2 days to install and index.

It was on my machine for 4.8 days.

It might be more help to others with less "content" on their machines. It does uninstall (Control Panel) cleanly and quickly, so it's safe enough to give it a try. (It is only available for WinXP and Win2K, and you must have Outlook or Outlook express installed to get previews of email.)

Of course once we all have a TB of stuff on our machines, something of this sort will be necessary; but I'm not ready for it with my piddly 0.15 TB of data now.

John