On May 5, 2013, at 9:26 AM, Turner <turner@xxxxxxxxx> wrote:
> In http://wiki.wireshark.org/Development/Translations we reached a conclusion that "Translating all the dissector output is practically impossible. But without this, it might not make real sense to translate other parts."
>
> By means of a RegExp search I figured out that there are roughly about 10000 string literals hard-coded in the dissector source files,
There are 126,824 named fields in a recent build of Wireshark.
*Every one of them* has a human-visible name field that shows up in the output of the dissectors; if you want to translate the dissector output, you would have to translate "Next server IP address" in BOOTP/DHCP, "Receive Buffer Posted" in SMB, etc., etc..
> but the majority of them aren't really user-visible strings that need to be translated.
All of those 126,824 fields are.
Many of those names are generated from ASN.1 source or other description languages.
Note also that Wireshark supports third-party plugins, and those aren't necessarily going to be translated into multiple languages.
There's also the WSGD plugin:
http://wsgd.free.fr
where translation would involve translating the text files that describe the protocols, which might be easier. (However, the syntax of the language *itself* is English-based - it's "uint1", not something involving an abbreviation of "entier", for example.)
And, in addition to that name field, there's also the field name, used in packet-matching expressions such as display filters, coloring rules, etc. and in the "-T fields" function in TShark. Would they need to be translated as well?
