Problem with diacritics

[+sernikk]

It's about time to write a topic about the problem with diacritics in the cache listing.

In Polish language there are some unique letters: ą, ż, ź, ś, ć, ę, ń, ó. Sadly, they are not showing up everywhere. There is no problem with using them in the main description, but in the hint or image description, there is.

 

What is happening?

When writing a hint on the edit cache page (lets use "ążźśćęńó" for test purposes) it looks like this:

 

But looking into the cache listing:

 

Yes. All letters except for the "ó" are converted into A-Z letters. The website is completely ignoring the dashes and little tails. Some words written that way sound completely different.

 

The same problem occurs in the "Edit Image" page (after uploading any photo to the cache listing). You just can't use these letters, and that's just sad. I guess it's about bad encoding and the core of the problem is somewhere deep in the web code. The exact same bug happens to be on the Waymarking.com page, but there it's even worse, because the website is getting rid of all these letters in the whole description. I can somewhat understand that Waymarking is not a focus for HQ (sad but understandable, I hope it changes soon), but geocaching.com in 2021 should allow to use all Polish letters, when it's even accepting emoji's (in hints, yeah!).

 

It would be really nice of you to fix it, not only for me, but for all my fellow Polish friends

Greetings

 

Edited January 11, 2021 by sernikk
correcting screenshot

[+ecanderson]

Since hints are managed by a simple ROT-13 substitution algorithm, which can only handle 26 characters, and expects an English alphabet, it comes as no surprise that it would mangle many other languages, including Polish.   That really is 'unfixable'.  However, it should not be a problem for image captioning.  That's something that should be looked into.

 

 

 

[+sernikk]

7 minutes ago, ecanderson said:

Since hints are managed by a simple ROT-13 substitution algorithm, which can only handle 26 characters, and expects an English alphabet, it comes as no surprise that it would mangle many other languages, including Polish.   That really is 'unfixable'.  However, it should not be a problem for image captioning.  That's something that should be looked into.

 

You're not exactly right It handles different languages easily. Example:

I used Spanish letters and symbols and here is the result:

 

They are showing up with no problems. Of course the ROT-13 is not going to work, but it shouldn't in this case so it's ok.

Edited January 11, 2021 by sernikk
typos

[+ecanderson]

I guess I'm confused.  How can there not be at least some issues when you attempt to use 13 x 2 substitutions for a 32 character alphabet?  I'm struggling with the math here.

 

[+mustakorppi]

2 hours ago, ecanderson said:

How can there not be at least some issues

If you type ä in a hint, it will not be "encrypted" by ROT13, but the letter is preserved and when the hint is decrypted, the hint reads like it should. That is, ROT13 treats it the same way it treats numbers in English text.

 

But if you type one of these Polish letters, apparently the original character is not preserved at all and is replaced even in the "decrypted" form.

[+sernikk]

Some time passed and the topic seems to be ignored, I hope it's not.

[+sernikk]

Yeah, it is..