Number of pages/paragraphs/words
Since we now have more accurate figures I have updated and extended this section. All numbers refer to the “core” documentation of perl, excluding the “core modules”.
The versions supported by this project (currently 5.6.2, 5.8.8, and 5.9.4) consist of 82,603 “strings” (or more appropriately called “paragraphs” in our case), alltogether. Although we use the same tools (gettext-tools, .po-files, etc.) as pure software translation projects there is a major difference: the length of such a “string”.
The strings (or paragraphs) we have do deal with are approximately 5-10 times longer than the common strings used in software (like “Open”, “Close”, or “Do you really want to quit the application?”).
An application working with the most recent version of all these docs would therefore have to be able to handle an equivalent of about 400,000 to 800,000 normal “strings” per natural language!
A counter example: OpenOfficeOrg has about 72,410 strings, Mozilla Firefox has about 39,957 (source: http://pootle.wordforge.org/). The help files for OoO 2.03, again, consist of 422,998 strings…
It is obvious that documentation, be it integrated into an application or standalone (like perldoc), needs a lot more resources (to store and to process) than simple dialogues.
The Rosetta project currently hosts about 732,581 strings alltogether (source: https://translations.launchpad.net/).
The core documentation (documents named of perl*.pod) as of version 5.8.8 consists of ca. 100 documents. This does not include platform-specific documents like perlmacosx.pod or perlos2.pod nor the core modules.
The overall amount of pages lies roughly between 1,500 to 2,000 pages of plain text (A4), or about 700,000 words. These figures can vary, depending on font size settings and on what your system considers a “word”.
From my experience the maximum amount that can be translated by one person lies between 100 and 150 pages per month. Depending on time, knowledge and experience these figures can greatly vary.
So the time one person needs to translate all the core documents lies somewhere between 10 and 20 months.
