A suggestion for Google

google suggestion

[heretic]

I can think of 3 different types of searches I do.

1. Search for arcane data. Any relevant response will do. (eg, "List of VI commands")

2. Search for news. (eg. Copyright info on google news page)

3. Search the WEB (not news sources) for more current information.

Now that google has that freshness robot, rather than going into advanced search and searching for last 3 months, an easier way to search for stuff 1-3 days would be cool.

Anyone share this thought? If not, how do you use google?

[mfagan]

You can search Google for things in the last 3 days. You just need to use their daterange: operator, which works so-so (they don't fully support it yet).

You can do this by searching directly (daterange:d1-d2, dates must be in the Julian caledar), or use an interface such as http://www.researchbuzz.com/toolbox/goofresh.shtml and also one that I've made.

[heretic]

ok, thanks.

[bcc1234]

That's one of the problems google is trying to solve. How to estimate a probability that a particlular page will have new content or new links and how soon.
Whenever they figure out how to "guess" that - we will all be able to enjoy truly fresh stuff from all over the web, not just a few lucky pages.

[heretic]

They could just implement a new html entity (or whatever it's called) or find an existing one, ask users to say how often they refresh their pages, check at the frequency they claim, if they're accurate consistently, crawl often, if not, penalize them. That should work pretty nicely.

[Grumpus]

That's one of the problems google is trying to solve.

I believe that this is not only the case, but that they're working very hard at it right now. In another thread, someone mentioned that freshbot was coming and hitting the same page several times in a row and/or over the course of a day. I can confirm this. This also seems to happen with "new" pages that get added to the index between crawls. When the page first shows up, it gets whacked a bunch of times the first day, and as the googlebot gets used to seeing it it backs off a bit and finally settles on once about every 3-5 days or not at all (except for the main crawl, I assume) depending upon the PR and fresh frequency of the page that links to it.

It seems like the bot is trying to learn a bit about each page and I'd guess it's trying to come up with a means of applying what it learns to my future additions.

On a site like mine, that's pretty easy to determine. I've really only got about 80 different "pages" that display different information depending upon the parameters passed to it. In general pages like X-Updates.asp and Y-Updates.asp tend to get updated several times a day as new content comes in. X-Details.asp and Y-Details.asp tend to stay pretty much the same once they get posted with updates happening infrequently, if at all. I'd like to think that Google will eventually identify that and I believe that what I'm seeing in my logs is evidence that they are attempting to do so.

G.