I am searching for the name Hishon. I have tried entering it as "Hishon", but I cannot prevent the retrieval of His Honour, his son and even Bishop. I presume that there must be a way of doing this, but where would I find such an instruction?
Perhaps that should be obvious, but it is not obvious to me.
There is lots of information on search methods provided here:
..... and more detailed information is provided here (including a direct answer to your question):
I would highly recommend becoming familiar with the different search methods, as it will make your searches much more productive.
In this case what you need is:
But you need to bear in mind the reason why Trove automatically does a 'fuzzy' search is so that you are more likely to find what you are after. A lot of the electronically translated text will not have been translated correctly, so the fuzzy search helps overcome that problem. The 'fulltext' search will miss all mistranslations and words like "Hishon's".
I have just had an experiment with the [ fulltext:hishon ] search and I see that it is still picking a lot of false positives, such as:
His hon. colleague (i.e His honourable colleague)
The reason for this is that in all these cases there is some spurious punctuation to be found between the "His" and the "hon." mistakenly inserted in place of the space by the electronic text translation, or the space was missed altogether. Punctuation is ignored during the indexing process for searching, so as a result "His.hon." or "His-hon." will be indexed as "Hishon".
You may be able slightly reduce the number of results by removing the ones that have been wrongly indexed because of a hyphen by doing this:
fulltext:hishon NOT fulltext:hon
But this would exclude all articles that included the abbreviation "hon." anywhere in the article. This works because the indexing process for hyphenated words, "His-hon." in this case, indexes both sides of the word as well as the whole joined word; so in this example it would have indexed (mistakenly as it turns out) "his","hon" and "hishon".
Another approach would be to rely on the 'relevance' ordering in the search results, rather than ordering by date. In this way the exact matches will be listed toward the beginning of the search results, however there are complications here too as several factors are considered in deciding relevance. For example, Advertising sections will be pushed toward the end of the search results even if they contain the correct spelling. This is all explained in the linked pages above.
Hope that helps.
Last edited by Spearth; 11-02-2012 at 05:03 PM.
Thanks, Spearth. I am penitent. I know that I should pay more attention to technique beyond the plain bread-and-butter work. But my brain turns to mush when confronted by on-screen instructions for anything to do with computers. (And Trove is not the only site on which I work, and still not generally the major one.) Reminds me of a child who asked how to spell a word and was told to look it up in a dictionary. "But how can I look up how to spell word when I don't know how to spell it?"
I do understand why Trove does fuzzy searches. That has saved my bacon often. But sometimes, as with Google, it tries to be too helpful. (I wish Google would stop forcing upon me what it thinks I ought to be looking for, instead of what I really want.)
I'll try to print out your instructions and experiment with them.
Sometime I find it useful to use Google in conjunction with Trove to find particular articles.
In this case, I'd go to Google, and in the search bar, type:-
This brings up all the articles including the word Hishon (a little over 600 articles). Although using this approach does mean you'll miss out on articles where the OCR has mis-read the name.
Thanks. I am having trouble with Google at present. When I type in something, Google automatically retrieves what it thinks I ought to want, without letting me finish my entry, and blocks me from entering anything more until it has had lunch. I think I need to have the program adjusted or re-entered. But right now I have too many other problems.
Even fulltext:hishon turns up over 8,000 results. (I have not restricted it to a time or place, as this is a rare instance when I am doing personal genealogy - my daughter-in-law's ancestors, not my own, so I am groping in the dark.) Patrick Mortimer Hishon was Chief Police Magistrate in Brisbane, and hundreds of trials where he presided as Police Magistrate are recorded. Great for tracking where he was, but a handicap if you want to find two other Patrick Hishons. I shall just have to keep at it industriously, correcting articles as they turn up.
Off and on, because this is the stuff of which madness is made.
In that case perhaps the following search may work better for you:
fulltext:hishon NOT (magistrate OR trial OR court OR "mr pm" OR "mr p m" OR "mr hishon pm" OR "mr hishon p m")
This search knocks the number of results back to 970.
This search will restrict the search to exact matches of "Hishon" (with the exception of mis-translations with punctuation as stated above), and also exclude any articles that also mention "magistrate" or "trial" or "court" or "Mr. P.M. Hishon" or "Mr. Hishon, P.M."). You may have to experiment with the words you include in brackets to see what gives you the most workable number of search results. Of course this would exclude any article that mentions that a particular Hishon had a relative who was a police magistrate, for example, but at least it might give you a good place to start.
If you are particularly looking for a Patrick other than the Police Magistrate then try this:
fulltext:hishon AND patrick NOT (magistrate OR trial OR court OR "mr pm" OR "mr p m" OR "mr hishon pm" OR "mr hishon p m")
Which gives only 78 search results; or:
"patrick hishon" NOT (magistrate OR trial OR court OR "mr pm" OR "mr p m" OR "mr hishon pm" OR "mr hishon p m")
Which gives just 3 search results!
There are always ways to refine the search.
Last edited by Spearth; 12-02-2012 at 05:12 PM.