There are a few potential gotchas with charsets. The main problem seems to derive from the fact your charset could be anything. Some are solvable but if you have mixed character sets then we're in a seriously unhappy place and I can't help you.

In some cases I've seen people whose charset is UTF-8 even though their database encoding is the default latin1_swedish_ci (which would appear to be the case). In these situations you should be able to migrate as is, validate using a UTF-8 capable editor that is in fact UTF-8 and then in the 1.5 side you select that it is UTF-8 and it shouldn't have a problem.

Other issues involve character munging along the way. This can some times be solved by on the 1.0 export configuring David Gal's UTF-8 guide if you don't have it already and then swapping back. This should create a UTF-8 dump file which you can then use to import into the 1.5 installer.

The last issue is perhaps the simpler: the database format doesn't match what you've got in there. Typically this is solved by exporting as per normal and then setting the correct character set of the data on the import routine. This will typically work well.

I'm sorry it hasn't worked for you, I'm testing a new release at the moment that will hopefully improve export performance, are you interested in some testing?

The Migrator has two places to translate. There is the language strings in the file in the 'language' folder which is used incidentally throughout the system and there are the files in the 'resources' folder which are for more template style translation and where the bulk text for a lot of items exists. I have responded to this effect on the bug report you opened on JoomlaCode and if after translating both of these places you are still missing strings, please list them on your existing ticket and I will update the migrator to fix this.