Thanks! and questions...

[BNDrinks]

Thanks for a cool search engine. We have integrated it into our newly re-launched site in two places:

http://www.barnonedrinks.com/drinks/search/
http://www.barnonedrinks.com/games/search/

We were able to make use of some of your more cool features, like OpenSearch integration. I was thrilled to find that I could just import a list of URLs and limit the search to a well-defined set of dynamic pages. And being able to specify the category in a meta tag was a lifesaver.

I have a couple questions arising from this implementation.

- Is there any way to remove the category display from each item in the results? The categories are displayed in our page titles, and are redundant in the search results.

- Is there any way to have a simple search form as the default, with an "Advanced" link including a URL parameter that enables the advanced search? I'd like to show users the simple search, but also allow them to switch to advanced mode if they need to.

Also, it would be nice to be able to define different URLs for OpenSearch based on combinations of categories, but I have no idea how you could present this functionality in a way that would make sense.

Thanks again!

Jason

[Ray]

- Is there any way to remove the category display from each item in the results? The categories are displayed in our page titles, and are redundant in the search results.

Sure. You could change the .category CSS class in your search_template.html to the following:

.category { color: #999999; display: none; }

This will hide the category text that appears next to your search result title (eg. "[Cocktails]"). You can do this to any of the other CSS elements to hide/display them as you need.

- Is there any way to have a simple search form as the default, with an "Advanced" link including a URL parameter that enables the advanced search? I'd like to show users the simple search, but also allow them to switch to advanced mode if they need to.

There isn't exactly that option, but one common approach is to have a separate search page which users are introduced to (eg. "mysearch.html") or you place in your menus (like you see we have at the top right corner of this very site) which can be easily accessed.

To do this, see this FAQ:
Q. How do I put search forms on pages besides the search page? (Or define my own search form?)

And then when they have submitted their search, and are taken to the search results page, the more advanced search form is displayed above the results for them to further narrow things down as needed.

Also, it would be nice to be able to define different URLs for OpenSearch based on combinations of categories, but I have no idea how you could present this functionality in a way that would make sense.

Not really sure how this would work either. I think we would need OpenSearch to allow categories to be specified as part of its standard for this to work? I don't know of anything like this at the moment.

[BNDrinks]

Sure. You could change the .category CSS class in your search_template.html to the following:

Awesome, thanks. I didn't even consider just supressing the display.

There isn't exactly that option, but one common approach is to...

Is there a chance you could add this as a request to your enhancements database? I'd really like to be able to keep the experience simple for the results as well, unless the user _really_ wants the advanced fields. What would be really cool would be a form template, where the implementor can define for each form type (simple, advanced) appears on each (categories as drop-down, categories as multi-select, max results, link to alternate form type, etc)

Not really sure how this would work either. I think we would need OpenSearch to allow categories to be specified as part of its standard for this to work? I don't know of anything like this at the moment.

I've defined multiple OpenSearch definition documents, with various categories defined statically as URL parameters in each. Unfortunately, the Zoom config only allows me to define one "rel" URL for the XML output. Right now, I have it set to the definition for all categories, but it's not entirely accurate that way. But, as I said, I have no idea how you could map between these definition documents and the parameters that the query is sent.

Anyway, thanks again; the speed of Zoom CGI is very impressive.

Oh, while I'm here. I'm looking at a FreeBSD 6.2 upgrade soon (we're at pair Networks too). Has the CGI binary been tested against this platform yet?

Jason

[Ray]

The CGI is currently built on FreeBSD 6.2-RELEASE-p1 so it should not be a problem.

[zoomnewbie]

We were able to make use of some of your more cool features, like OpenSearch integration. I was thrilled to find that I could just import a list of URLs and limit the search to a well-defined set of dynamic pages.

Limiting the search to a specific set of pages on the website would indeed be something I could use on my website, but I thought there was no way to do that in Zoom. My understanding is that one could "exclude" certain pages from a search within a URL, but not make a list of only certain pages within that URL for Zoom to search.

How do you limit Zoom's search engine to just certain pages within a website without tediously excluding the many ones you do not want searched and indexed? That would be an extremely useful option for me as I do not want the entire website searched which would yield meaningless results.

Any other way to make Zoom only search certain pages within the site?

[Ray]

Limiting the search to a specific set of pages on the website would indeed be something I could use on my website, but I thought there was no way to do that in Zoom. My understanding is that one could "exclude" certain pages from a search within a URL, but not make a list of only certain pages within that URL for Zoom to search.

Actually, the Categories option in Zoom should cater for this. Take a look at chapter 2.5 of the Users Guide for information on this.

How do you limit Zoom's search engine to just certain pages within a website without tediously excluding the many ones you do not want searched and indexed? That would be an extremely useful option for me as I do not want the entire website searched which would yield meaningless results.

The categories option allow the end user to select a specific (pre-determined) group of files to be searched. But if you have files which you want to permanently exclude from all searches, then you should omit them from indexing in the first place by adding entries in the "Skip Options" tab of the Configuration window. Note that the skip pages do not require each URL to be entered, just a pattern which it can match by (e.g. a skip entry of "/private/" will skip all files in any folder named "private").

See chapter 2.2.4 of the Users Guide for more information on Skip Options.

[zoomnewbie]

No, I do not think that is what I was looking for. What I am looking for is
1. say you have a site with 100 html files.
2. you want Zoom to include in its search only 20 of them.
3. Rather than write the individual names of the 80 html files you wish to "exclude" (which is what Zoom's skip option now permits), it would be nicer to be able to instead write the names of the 20 html files you do want to "include".

I assume one way around this might be to move these 20 files into their own separate directory, but that would disrupt the site's internal relative links and cause a lot of rewriting. So having a place to list "include" the following specific files would be nice rather than the present options that include all files that share similar extensions.

[Ray]

OK, I see what you mean now. There is an easy way to do this, actually. You simply need to specify each page as a start point.

In Spider Mode, click "More". Select "Add", and enter in one of the twenty URLs you wish to index. Then select the "Index single page only" option, and click OK.

This start point will only index one single file, and not follow links to any other pages. You can repeat this step to add the other 20 files, each as an individual start point.

If you find that too tedious, or you have a longer list to manage, use the "Import" feature, and create the list beforehand in a text file (make sure to specify the INDEX_ONLY option. See the Users Guide for details).

[zoomnewbie]

Thank you. That indeed is what I wanted and a convenient way to do it. Thank you. I think I should have read the users guide more carefully before asking, but I only rapidly went through it. (You know how some computer users are .......).
Thank you.

[Ray]

No problem. There's a fair bit in there, and it's quite easy to miss some features. As nice as it would be for everyone to read the Users Guide thoroughly, we know that's never going to be the reality

[BNDrinks]

Sorry, my server's hard drive crashed so was a bit distracted...

Yes, that's the solution I was using, and I agree that there are lots of cool features to find in Zoom.

I was using the import feature with INDEX_ONLY (or something like that). It worked quite well with over 8000 pages in a single profile, though it took a while to come back on my creaky old Athlon 900. I also used category meta tags in the pages within this set (dynamically generated), so that these well-known pages could be broken down into well-known categories

Not a typical use for a search engine, but for database-based websites it's a great way of tailoring your search.

What I would really like would be a way of combining multiple search results in a single search. So I could have four individual indices (one for my static pages, one for each of the three dynamic sections), and then have a site-wide search that uses all of them. I'd imagine that this would not be all that efficient though, so I think I'm stuck maintaining a meta-index that includes all of the sections.

Jason

[David]

You could use MasterNode to combine results from multiple search engines / sets of index files.

[BNDrinks]

That looks like exactly what I need but... it looks like it would require me to be on a Windows box for the main search form? The indexing part isn't a problem, but my only webserver is FreeBSD.