Page 1 of 1

The following words in your search query were ignored because they are too common words: port serial com

Posted: Tue Dec 31, 2019 1:09 am
by Andy2No
I meant to mention this in the old forum, but it disappeared too quickly...

The forum search has a habit of rejecting the very words I want to search on, because it says they are too common.

If it was just words like "the", "it", "why", that would make sense, but it's technical, relevant keywords, most of the time, like "port", "serial", "com"... and many others.

If the forum is to avoid having people ask the same questions over and over, even though they've been answered many times before, I think it would help if that could be turned off, or limited to non-technical words, so people could actually find relevant topics by searching, and not have to ask someone, in a new thread.

I'm now off to start a new thread...

Re: The following words in your search query were ignored because they are too common words: port serial com

Posted: Tue Dec 31, 2019 3:00 am
by fredbox
In the old forum it was mentioned several times. The mentioner was usually advised to try advanced searches with Google.

The workaround is to use an asterisk. Serial is too common but Serial* works just fine. Port is too common but Port* returns many matches.

Re: The following words in your search query were ignored because they are too common words: port serial com

Posted: Tue Dec 31, 2019 6:17 pm
by Andy2No
Thanks for the workaround.

The real flaw in rejecting words that are too common is that this sort of forum engine still does it for multiple words - even where the combination might be a lot less common.

Yesterday, I searched for "virtual com port serial" (without quotes), which got reduced to "virtual" - not much use at all. I see today it returns just one result, which was me mentioning the problem, in an other thread... I'm not sure if that means something has been done to change it, or not. I didn't get the same message, so maybe it has.

Re: The following words in your search query were ignored because they are too common words: port serial com

Posted: Thu Jan 02, 2020 9:13 am
by fpiSTM
Currently, I'v made no change on how phpBB manage search.
I'm open to all suggestion to optimize it anyway I guess the default should be fine regarding phpBB legacy.

Re: The following words in your search query were ignored because they are too common words: port serial com

Posted: Fri Jan 03, 2020 1:32 pm
by racemaniac
fredbox wrote: Tue Dec 31, 2019 3:00 am In the old forum it was mentioned several times. The mentioner was usually advised to try advanced searches with Google.

The workaround is to use an asterisk. Serial is too common but Serial* works just fine. Port is too common but Port* returns many matches.
Annoyingly, since this forum has basically been reset, most google results won't work >_<

Re: The following words in your search query were ignored because they are too common words: port serial com

Posted: Fri Jan 03, 2020 2:10 pm
by ag123
i tried serial*
search.php?keywords=serial*
that seem to work around the issue, thanks for the tip!
google:
serial site:www.stm32duino.com

interestingly returns results from the old stm32duino forum. i'd guess these days the algorithms and indexing is more particular than it used to be.
and perhaps crawling and indexing isn't as frequent as in the past due to the already huge data size that's probably sitting in google today
anyone tried bing and the rest?

Re: The following words in your search query were ignored because they are too common words: port serial com

Posted: Mon Jan 06, 2020 4:07 pm
by Andy2No
fpiSTM wrote: Thu Jan 02, 2020 9:13 am Currently, I'v made no change on how phpBB manage search.
I'm open to all suggestion to optimize it anyway I guess the default should be fine regarding phpBB legacy.
I used to use a radio control forum which used phpBB, and looked similar to this one. The search got progressively worse, until it was basically unusable.

The algorithm appears to take "most common" words from the forum database, because the very words you are most likely to need to search on get classed as being too common - but are not particularly common words in the English language.

I think the best search is one that does a strict "AND" (all keywords present) search. If people put in too many words, they get very few results, so they learn not to.

Failing that, if you can turn off the "too common" rule, that would be a good start.

Re: The following words in your search query were ignored because they are too common words: port serial com

Posted: Tue Jan 07, 2020 2:37 pm
by fpiSTM
I've modified the threshold. it seems more permissive.

Re: The following words in your search query were ignored because they are too common words: port serial com

Posted: Tue Jan 07, 2020 6:49 pm
by Andy2No
That sounds promising. Thanks, fpiSTM.