Tried searching for 'Another Armitage' in the search for an author area. This is a valid author's name. The box isn't long enough to handle an author's name of this length and will notwork on part of the name. I had to resort to manual searching.
Tried searching for 'Another Armitage' in the search for an author area. This is a valid author's name. The box isn't long enough to handle an author's name of this length and will notwork on part of the name. I had to resort to manual searching.
Searches for "another" or "anoth" both work.
I thought in might be the combination of the full length search sting and having a space in the string, but a test using author "A Purvversion" works fine.
I can reproduce the failed search by the OP.
That search box only accepts 10 characters which should be sufficient to get a very limited set of results containing the author you are looking for. The good thing is that you can search any part of an author name. So for example searching for "sky" or "ncle" will have "Uncle Sky" among the results. Very handy if you don't know the exact name, I use it a lot.
Yes, but it's still odd that it fails for this one author if the partial search string is at the length limit.
A problem I have had with Author Searches is that the non-letter characters have to be entered exactly, where these are the least memorable parts of the name, i.e. the names
A Name
A.Name
A. Name
A-Name
A_Name
would each need an exact match, whereas I would have hoped that each would match any and also AName. Searching on 'contains' Name would work, but could easily generate a list of too many matches, as would searching on 'starts' A. The obvious way round this is to remove all non-letters from the strings before comparison, however that exposes another bug which I've seen on the 'search title' field in Advanced search.
The latter bug is that accented characters have to be matched exactly - try EB's 'Same sex marriage débâcle'. The matching should be done between character class equivalents so â matches a etc. Care should be taken in the 'eliminate non-letters' above that accented characters are not replaced, or are replaced by non-accented versions. I know Bash has functionality to match with character class equivalent, but I don't know about replacing with the base character.
A problem I have had with Author Searches is that the non-letter characters have to be entered exactly,
Doesn't the author's name contain an underscore?
AJ
The latter bug is that accented characters have to be matched exactly - try EB's 'Same sex marriage débâcle'. The matching should be done between character class equivalents so â matches a etc. Care should be taken in the 'eliminate non-letters' above that accented characters are not replaced, or are replaced by non-accented versions. I know Bash has functionality to match with character class equivalent, but I don't know about replacing with the base character.
It's called string normalization and I have yet to find the perfect normalizer. I work a lot with diacritics, it's part of my job. The most frustrating problem is that you can't just convert a diacritic to the base letter because it's language dependent. For example the German Jäger should be converted to Jaeger, not Jager where in other languages the conversion to the base letter a is correct. There are more of such 'rules' which are language dependent. The perfect function to normalize a string has a language parameter to enable parsing to the correct base letter(s).