Pound Sign character
Administrator
Joined: Oct 2000
Posts: 20,274
Likes: 4
From: Toronto, Canada
character coding is a nasty business but here is the gist.
UTF-8 is a character set encoding. ISO 8895-1 is a different encoding. English/latin letters and number are the same in both encoding systems so "hello world" is the same in both. Greek, Cyrillic, Arabic and Asian characters are different in each encoding system and in some cases not supported at all. UTF-8 supports all characters from all languages. ISO-8895-1 supports only 1 byte per character languages so no Japanese or Chinese or anything which requires more than a single byte per character.
When the site was started 8 years ago we setup the database and the software to encode everything in ISO 8895-1 character. All of the data is is stored in that character set.
When I changed the server software a couple of days ago it was set to default settings which is UTF-8. So anything you posted in the past few days was sent to the server as UTF-8 characters. However, the software assumed it was ISO and as far as it was concerned was.
So long as the server was sending it back to your browser marked as UTF-8 everything was cool and nothing was lost in the translation. However, the server was telling your browser than everything post last week, last month and last year was also UTF-8 which it is not. So all the posts from the past are illegible because your browser it being told they are in UTF-8 format and they are not. For English it doesn't matter because all the characters are the same.
I fixed the server setting and changed the charset back to ISO-8859-1. Now all the posts from the past are correctly identified as ISO characters and are legible. However, posts made in the last few days which were sent to the server as UTF-8 are now being send back to you marked as ISO 8859-1 so now they are illegible as far as non-English characters go.
Apologies to the Greek members for the mixup. Unfortunately you'll need to go back and edit your posts from the past few days and convert them. There is no other way.
UTF-8 is a character set encoding. ISO 8895-1 is a different encoding. English/latin letters and number are the same in both encoding systems so "hello world" is the same in both. Greek, Cyrillic, Arabic and Asian characters are different in each encoding system and in some cases not supported at all. UTF-8 supports all characters from all languages. ISO-8895-1 supports only 1 byte per character languages so no Japanese or Chinese or anything which requires more than a single byte per character.
When the site was started 8 years ago we setup the database and the software to encode everything in ISO 8895-1 character. All of the data is is stored in that character set.
When I changed the server software a couple of days ago it was set to default settings which is UTF-8. So anything you posted in the past few days was sent to the server as UTF-8 characters. However, the software assumed it was ISO and as far as it was concerned was.
So long as the server was sending it back to your browser marked as UTF-8 everything was cool and nothing was lost in the translation. However, the server was telling your browser than everything post last week, last month and last year was also UTF-8 which it is not. So all the posts from the past are illegible because your browser it being told they are in UTF-8 format and they are not. For English it doesn't matter because all the characters are the same.
I fixed the server setting and changed the charset back to ISO-8859-1. Now all the posts from the past are correctly identified as ISO characters and are legible. However, posts made in the last few days which were sent to the server as UTF-8 are now being send back to you marked as ISO 8859-1 so now they are illegible as far as non-English characters go.
Apologies to the Greek members for the mixup. Unfortunately you'll need to go back and edit your posts from the past few days and convert them. There is no other way.
Registered User
Joined: Dec 2004
Posts: 30,032
Likes: 2
From: Athens, Greece, Europe, Earth, Universe
Since yesterday I went through all the posts in our greek community and edit them from "UTF-8" into "Western European (ISO)". It was not easy but it is sorted now and everything is cleared. 
Thread
Thread Starter
Forum
Replies
Last Post