The following suggestions are often presented as being solutions:
Why not use a decode table as a pointer into Telnet client?
or
Our browser has a decode table, therefore we have solved the problem this way.
or
We have proxy channel for that Internet service, it decodes automatically.
If you consider this a little deeper, you will see that it will not always work... If you utilize this solution, a decode table or proxy would have to exist in each and every interactive Internet service, i.e. telnet, news, gopher, finger, WAIS, HTTP, FTP, etc. Even if you cover all standard Internet services by proxy channels or decode tables, you will be hitted by any newly appeared service, f.e. various multi-user conferencing tools. Lots of such services grows up nowadays.
The disadvantages of decode tables method soon become very obvious, because most of the existing applications don't support decode tables. That is, you would have to restrict the user to a your particular WWW browser with a particular decode table, however many very good browsers are in existence, so it would be a very unwise and confining restriction. Of course you could contact each and every software author and try to convince him to support your decode table, but ... good luck with doing that!
The disadvantages of proxy method become obvious too, because you can cover only a few old, well-known and stable Internet services and can't deal with others. You need to program and debug a separate proxy for each service. There are some bottlenecks even for known services' proxies such as HTTP, e.g. most proxies do not remove <META ... charset=...> fragments from <HEAD> section. It won't cause any inconvenience with browsers that follow standards (HTTP header charset must override META charset), but this can badly damage decoding for broken browsers that act otherwise, e.g. Mozilla and MS Explorer.
I think the easiest and the most reliable way to solve this problem is to have two character sets installed, CP1251 and KOI8-R, and then switch between them. People who disagree with this approach often say:
But you need a program to convert between two charsets!
Well, of course you do! But you need such a program in either case, i.e. you receive a binary attachment which is text in KOI8-R: no decode table would help you in this case. But in my variant (i.e. if you have KOI8-R fonts installed too) you can read, edit, cut & paste such documents easily!
Better English variant of this section provided by Steve Van Osdell <svanos@cdc.net>, thanx to him.