Otdfbt

I have just become DBA for a database that support call monitoring software attached to our PBX. The software was developed in Italy and all objects (Tables, Columns, Procedures, Views etc) are named with Italian names e.g. the Chiamate table holds the Call data.

Can anyone suggest how I can build up a table of translations - a simple two column Italian to English table that I can use to reference which columns,tables I should use to report from - until I get to learn Italian!!??

I had an idea that it may be possible to use an online translation service?

any ideas gratefully recieved.

Jonathan

Here is a good Italian-English online dictionary.

You could also build a list of all table names, and let it run through Microsoft's translator, which accepts whole texts:

 select name from sysobjects where type = 'u' -- = user tables

You could do the same for column names joining syscolumns etc., for example this way:

select TABLE_NAME, COLUMN_NAME
from INFORMATION_SCHEMA.COLUMNS
order by TABLE_NAME, ORDINAL_POSITION

Another additional tip: you could create SQL Views using the translated table and column names and work with the views.

Buona fortuna!

Update: in a comment to this answer, Fatherjack writes:

I have [written] a small .net app
that reads in a text file (each line
is a table name . column name), passes
it to www.microsofttranslator.com and
writes out the response to a new text
file. using
api.microsofttranslator.com/V1/SOAP.svc.

HOW TO's found here, examples
here

Get a good Italian to English dictionary and do it manually - the results from babelfish or a mechanical translation service will probably be slightly nonsensical and need manuan intervention anyway.

As an example, this real estate listing ...

TRILOCALE IN PICCOLO CONDOMINIO CON
POCHE UNITA' ABITATIVE. L'APPARTAMENTO
è SITO AL PRIMO PIANO E SI COMPONE DI:
INGRESSO, CUCINA ABITABILE, SOGGIORNO,
2 CAMERE, SERVIZIO, 2 BALCONI E BOX.
LIBERO SUBITO, NON ARREDATO, ZONA
TRANQUILLA.

... gets translated by babelfish to:

TRILOCALE IN LITTLE ONE CONDOMINIO
WITH LITTLE UNITA' INHABITED. L'
APARTMENT is SITUATED FIRST AL SLOWLY
AND IT IS MADE UP OF: INCOME,
HABITABLE KITCHEN, STAY, 2 ROOMS,
SERVICE, 2 BALCONIES AND BOX. FREE
QUICKLY, NOT FURNISHED, CALM ZONE.

Real estate listings are fairly idiomatic, but then so are database schemas. To take the above example, unless you knew that an Ingresso was a porch, a Soggiorno was a lounge or a Cucina Abitabile was an open-plan kitchen you wouldn't get much sense from this translation. Also, some of the jargon is idiomatic Italian - a Trilocale (roughly: three spaces) is a two-bedroom apartment. The living area and bedrooms figure in the count of rooms but bathrooms and suchlike do not. Mechanical translators are typically going to do poorly unless they are aware of these idiomatic rules.

A database schema will have highly idiomatic wordings in it, and possibly lots of abbreviations. SAP R/3 is famous for its cryptic database schema where the table names and field names are based on abbreviations of German words. It is difficult to understand if you are not a fluent German speaker (and not necessarily that easy if you are). Mechanical translation is unlikely to help much with this.

You will find that Italian has many root words in common with English and with a little tourist-level working knowledge of the language it is not hard to work out what a written passage means. If the tables have a lot of abbreviated words making up the column names you may still have trouble. I would consider taking up splattne's offer to help with the transalation if you run into this type of problem.

Translating between languages is highly dependent on context. As a simultaneous interpreter, my first question on a job is, "Di che cosa si tratta?" "What is this about?" So, the "meaning of a (a word in a) column would depend on (the word in) column preceding it, and possibly following it. Google translate takes this into account to some degree (the sliding window approach). Since here the context seems somewhat restricted, recursion may be simpler. The lookup should include as much "common information", context, as possible. That information might be obtained by passing some of the free-text through some classifier program.