I have recently fin­ished trans­lat­ing Gramps 3.4 to Por­tuguese (Europe), some­thing I had star­ted some time ago but let it bit rot a bit; this revi­sion is quite extens­ive and I tried to review every exist­ing sen­tences and make them more harmonious.

Here is the po file for download.

The key points of my trans­la­tion (which I had already stated in my Gramps pro­ject wiki page) were from the start:

1. Increase con­sist­ency of the trans­la­tion, by review­ing all the sen­tences after decid­ing for a par­tic­u­lar construct
2. Use ver­nacu­lar ter­min­o­logy; there is a nat­ural tend­ency to trans­late “word by word” and dis­miss exist­ing expres­sions because they are not so well known (since they are part of a more or less tech­nical lan­guage). Using them both increases the qual­ity of the trans­la­tion and avoids import­ing terms which have a per­fectly suit­able local alternative
3. Avoid chan­ging things because of slight pref­er­ence; this is mainly dir­ec­ted at the dif­fer­ences in taste between Por­tuguese in Por­tugal and Brasil, which reflect in the order of the ele­ments and pre­ferred used of some syn­onyms; my approach is that it is waste­ful to change things that are per­fectly reas­on­able  since that decreases the con­sist­ency of trans­la­tions when import­ing new additions.
4. Use the latest agreed ortho­graphy, some­thing I par­tic­u­larly dis­like as I had the oppor­tun­ity to share in my Guarda Gozosa post
5. Build a decent style guide for trans­lat­ing gene­a­logy pro­grams to Por­tuguese; Gramps is an excel­lent pro­gram, but this effort is actu­ally use­ful bey­ond Gramps since the same chal­lenges apply to any other gene­a­lo­gical application.

I’m not entirely happy with the trans­la­tion still, but it’s 100% com­plete and it’s bet­ter to refine it from there than to have it only half done, even because now it is per­fectly pos­sible to use it as the main lan­guage in Gramps (I’m using it that way), which will nat­ur­ally lead to more errors being corrected.

My object­ive is to con­tinue to explore the most adequate way of incor­por­at­ing por­tuguese gene­a­lo­gical terms and praxis into Gramps, which is by now in my opin­ion one of the best if not the isol­ated win­ner in terms of sup­port­ing mul­tiple sur­names. The whole open nature of Gramps really shines here: it star­ted as a bug request, pro­gressed into a GEPS, an was finally imple­men­ted. This func­tion­al­ity is to my know­ledge unique in the way it com­pletely sup­ports mul­tiple sur­names, includ­ing tra­cing them through­out the gen­er­a­tions even when they are passed in non-obvious ways.

Edit­ing po files was some­thing that was made much easier by Emacs; ini­tially when it was mostly about non-complex data entry I used gtrans­lator, poedit and vir­tall. Each one has some advant­ages and draw­backs, and per­haps sur­pris­ingly I ended up using gtrans­lator the most because I wasn’t able to make the oth­ers show me how a spe­cific entry was trans­lated in three other lan­guages — some­thing I like to do to use as ref­er­ence. Emacs po-mode works quite well and I was able to appre­ci­ate it more as I star­ted to under­stand more about the gettext-based trans­lat­ing cycle. At the end, when I was detect­ing small dis­crep­an­cies between the ori­ginal sen­tence and the trans­la­tion (say, a rogue final space) Emacs reg­exp func­tions coupled with a key­board macro made the pro­cess close to auto­matic, allow­ing me to retouch ~500 entries in minutes instead of hours.

There is still room for improve­ment though:

1. I would like to see the aux­il­i­ary trans­la­tions in win­dows, i.e. not cyc­ling through the sev­eral files but hav­ing them as sep­ar­ate win­dows that would be auto­mat­ic­ally updated
2. I tried and failed to come up with a way to auto­mat­ic­ally detect and fix the cases where the trans­la­tion misses a final period; this would be extremely use­ful but I wasn’t able to make mul­til­ine reg­exp match­ing work with my use case.

If you’re inter­ested in gene­a­logy do give Gramps a go!

PS: I have also recently looked at a new pro­gram I found, Ances­tris. It has a com­pletely dif­fer­ent approach from Gramps and derives from it the biggest advant­ages and draw­backs: it uses GEDCOM as the nat­ive format, which makes it 100% GEDCOM com­pat­ible. Addi­tion­ally it runs every­where (it’s Java-based) and the data­base is a single .ged file (that can be accessed over the net­work if needed). Well worth a look!