Talk:Isearch

Isearch

 * should not be considered merely an "experiment". It is still routinely distributed as part of the ports collection of BSD (textproc/isearch).
 * It has had a large number of users and has been at the search core of many well known sites (including the Open Directory Project which started as NewHoo).
 * Is hidden in several commercial software packages.
 * It still has a significant scientific installed base of users in the Geospatial Clearinghouse. Most sites under http://registry.gsdi.org/serverstatus/ are running the public Isearch code! http://registry.gsdi.org/serverstatus/ shows 427 nodes in the network.
 * It continues to be developed and supported.
 * It forms the basis of numerous split-off developments including the patent search of the WIPO (World Intellectual Property Organization) and the US PTO. —Preceding unsigned comment added by Edward Zimmermann (talk • contribs) 16 November, 2007
 * Comment. Isearch may be used in a number of places, and some people may like it, but it needs to satisfy Wikipedia's standards of notability for us to keep an article on it. Someone added it to List of search engines, where it does not seem to belong, since the other items in that list are functioning web sites that will accept search requests. To justify an article on Isearch, we would need to see reliable sources, independent of the authors of the software, testifying to its importance. Usually that implies articles in edited print publications, or articles in edited web sites that maintain a reporting staff.  At present, the article does not include any such sources. EdJohnston (talk) 23:59, 16 November 2007 (UTC)

I've not included references to them but references that include links or other references to services current and historical. To "functioning web sites that will accept search requests": That is a service and not an engine. Isearch (as well as all, I think, of the engines in that list) are solutions and not services but, for the most part, software packages, programs or libraries. And to "functioning web sites". There are still 1000s of functioning "web sites" that use the public Isearch code for search services (or derivative works) but I don't see "functioning" as a necessary condition of noteworthiness (either today or in the past). Isearch, for example, was the software behind "Internet AIDS Patent Library": http://www.aegis.com/news/ads/1994/AD941940.html in 1994 (this was at the highest level with Ron Brown, then United States Secretary of Commerce at the press conference doing the demonstration See: http://www.hpcwire.com/archives/3149.html) and was later followed by the public search service of the United States Patent and Trademark Office. There is a lot more! You judge? Insignificant "experiment"?

Edward Zimmermann (talk) 10:06, 18 November 2007 (UTC)
 * Your expansion of the article is impressive, but there are no inline citations to reliable sources, so it is hard to be sure that the content of the article can be confirmed from the sources given. (A skeptical reader might think that you were putting statements in based on your personal knowledge). Also there is not really any sourced evaluation provided. There are no comparisons of Isearch against rival methods, so it's hard to know whether Isearch is obsolete. EdJohnston (talk) 14:26, 18 November 2007 (UTC)

What kind of citations would you like. There are LOADs out there. Evaluations? OK.. I'll add a few links to comparison evaluations I've seen over the years (if I can find them)..

Edward Zimmermann (talk) 15:51, 18 November 2007 (UTC)

In term of performance? I personally think the public Isearch code is very limited. Baeza-Yates even wrote a few articles discussing and addressing the problems of those algorithms. Put simply: too dependent upon file I/O (and opening individual files in Unix is very very expensive!). There are, however, large numbers of very significant applications that have important information but not large amounts to store in their individual indexes. See the Geospatial Clearinghouse activity where most of the nodes, I think, have relatively small numbers of records. Significant in total but not terribly much in each node. I've included a link to a presentation on its relevance to African development. Recall this is a distributed model-- today the fashion might be to nearly call it P2P--- and not a centralized monolithic design for search. In that perspective, I think, the issue of indexing performance and speed is moot. More significant (from a strategic view) are its special features especially its geospatial search support and cost (FREE).

Edward Zimmermann (talk) 16:27, 18 November 2007 (UTC) While I included Eric's comments-- and I plan on including a few more comparisons once I find them (there were quite a few)--- I would not give them any special credence in the text since 1) Isearch/Isite is still supported and being developed. I just checked the most recent public release was 4 Sept 2007 2) The historical CGI interface he is referring to was, according to documentation, never intended for real use but as a working example. The intended interface was via the Z39.50 protocol (most often the Isite ISO 23950 stack or via the YAZ library as-per the ASF distribution) and most significantly 3) I think he looked at a very old release tree that probably dates from the CNIDR days which effectively ended, as I mention in the text, no later than 1998 when Jim Fulton (the project leader and, I suspect, the instrumental figure behind the creation of CNIDR) left to head up the patent search group at the WIPO in Geneva. The ASF project of Dept. of Commerce issued a ready-to-run CD and FGDC distributes also special versions for out of the box creation of Clearinghouse nodes. If I recall there is a lot of documentation at FDGC. I'd link into Eliot's gils.net site but its down today (he "officially" retired from the USGS 1 1/2 years ago so I don't know its status). Edward Zimmermann (talk) 18:46, 18 November 2007 (UTC)
 * The document by Eric Morgan that you recently added includes this comment about Isearch: "Unfortunately, Isite/Isearch no longer seems to be supported and the documentation is weak. While it comes with a CGI interface and is easily installed, the user interface is difficult to understand and needs a lot of tweaking before it can be called usable by today's standards." While this sounds like a negative review, our articles are supposed to be neutral, so do you think that this information could be added to the text of the article? (Attributed as Morgan's opinion, of course). EdJohnston (talk) 17:01, 18 November 2007 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 2 external links on Isearch. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20060315101429/http://www.etymon.com/tr.html to http://www.etymon.com/tr.html
 * Added archive https://web.archive.org/web/20070618105456/http://www.ibu.de/node/52 to http://www.ibu.de/node/52

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 04:30, 17 November 2017 (UTC)