Non english letters

Reads 2333 • Replies 34 • Started Tuesday, October 17, 2017 3:42:51 PM CT

The forums you're viewing are the static, archived version. You won't be able to post or reply here.
Our new, modern forums are here:
RateBeer Forums

Thread Frozen
 
Koelschtrinker
beers 28840 º places 22 º 15:42 Tue 10/17/2017

So, we are having the letter problem since a longer time know. A search of Mühlen Kölsch e.g. brings no result ( https://www.ratebeer.com/beer/mhlen-klsch/7778/ ) and I have to edit every upload from my mobile and replace question marks with the actual letter.

When do we get this fixed? I'm not too much into Web Server, but isn't this something like "Settings --> Use UTF-8" and it would be done?

 
Bitterbill
beers 2825 º places 25 º 15:51 Tue 10/17/2017

Your link brought up the right result. Using website on my mobile.

 
Koelschtrinker
beers 28840 º places 22 º 16:02 Tue 10/17/2017

Sure, link works. but search for "Mühlen Kölsch" or just "Kölsch", that will bring up around 20 results while there are way more than 100 beers on RB with Kölsch in it

 
FatPhil
beers 21223 º places 832 º 16:08 Tue 10/17/2017

When searching for foreignish words - drop the dots:
""
Results for muhlen kolsch

Show exact matches only

beers
Mühlen Kölsch
""

That doesn't always work, recently searches for P6hjala (yeah, I can't type o-twiddle) would find 9 of their beers, and searches for Pohjala would find all of their other beers, with no overlap. So no one method worked for all beers.

Apart from just doing your rating on unta!:#@^Q*#*!(NO CARRIER

 
Koelschtrinker
beers 28840 º places 22 º 16:10 Tue 10/17/2017

Originally posted by FatPhil
When searching for foreignish words - drop the dots:
"
Results for muhlen kolsch

Show exact matches only

beers
Mühlen Kölsch
"

That doesn't always work, recently searches for P6hjala (yeah, I can't type o-twiddle) would find 9 of their beers, and searches for Pohjala would find all of their other beers, with no overlap. So no one method worked for all beers.

Apart from just doing your rating on unta!:#@^Q*#*!(NO CARRIER


Sure, dropping the points works, but thats not a solution. How hard can it be to fixed this problem (which only exists since a few weeks [meh, probably months])? It just sucks.

 
Bitterbill
beers 2825 º places 25 º 16:26 Tue 10/17/2017

Originally posted by Koelschtrinker
Sure, link works. but search for "Mühlen Kölsch" or just "Kölsch", that will bring up around 20 results while there are way more than 100 beers on RB with Kölsch in it


Why insist on using the umlauts?

 
joet
admin
beers 2898 º places 125 º 17:14 Tue 10/17/2017

Originally posted by Koelschtrinker
When do we get this fixed? I'm not too much into Web Server, but isn't this something like "Settings --> Use UTF-8" and it would be done?


"COMPUTER! ENHANCE!"

if things were only this easy. 😅

 
joet
admin
beers 2898 º places 125 º 17:16 Tue 10/17/2017

Obviously I have some work to do. Just got done with finishing up Swedish places... will get to this soon.

 
joet
admin
beers 2898 º places 125 º 17:22 Tue 10/17/2017

OK. done. And beer search is improved. To explain a bit... we have a whole bunch of data and a long history about how this data was encoded and stored in the database. Some are stored as HTML entities, some as alternately translated Unicode, some as native UTF-8. Microsoft has no silver bullet and developers have implemented dozens of different methods for dealing with character issues. On top of this, Microsoft has separate methods and data types for storing unicode versus non-unicode. After we have some stability on this front, which we hope to soon, we can begin the long (in computer scale time) process of translation.

At the same time we are translating scripts, which have server default encoding, file encoding, application encoding, and headers for browser encoding. Sometimes one script will reference a dozen files, all of which must have the correct encodings. We automate as many changes as possible but RateBeer has a whole lot of files, and being built over 17 years, has many different coding styles and practices -- many of which my own over the years. This complicates automated methods.

 
LazyPyro
beers 5280 º places 59 º 18:28 Tue 10/17/2017

Microsoft... that explains it. Notoriously inconsistent with character encoding.

Romanian is another charset that needs fixing. Currently some beers are impossible to find unless you know the brewery name (not always obvious for macros). For example Timișoreana. Searching either Timișoreana with the correct s or Timisoreana... neither return any results at all. The only way to get to them is a Google search, or by going to the Ursus brewery page first assuming users know it's brewed there.

 
LazyPyro
beers 5280 º places 59 º 18:54 Tue 10/17/2017

Oh and btw it's hard to report issues like this through the feedback page because I can see from the email confirmation that the charset is not encoded properly, even standard quote marks become gibberish. Line breaks are also removed making readability difficult for long reports.