I'd like to know if anyone has worked on Silk Test International. I want more information on Internationalization/Localization Testing using silk test international. Anyone worked/working on Internationlisation./..Please Help. Is it sufficient to record the script in one langauage and run it for different languages ?
good point sir!!!!!!!!!!!!!!!!
we are in the process of such a transition.
We recorded all scripts on english locale and now the time has come to run these scripts on german and swedish locale for the same product.
following stpe was taken to internationalize the scripts.
1) replace the window multitags with indexes
(e.g. : tag #1 instead of
tag "my window" )
(e.g. htmllink GO
but by and large ...the hardcoded strings in the t file had to be manually replaced
any better suggestions or are we destined to do this tedious job??
First of all. SilkTestI is only necessary if you are testing on double-byte o/s's. Otherwise, the regular SilkTest will work just fine on other languages.
When possible, we declare everything using control ID's only. When it's not possible, we use string variables for the tag. The string variables are defined in an include file with the english translation. We then put the foreign translation of the variables in a text file and then parse out the text file in ScriptEnter if the o/s is a non-english language. So if I'm testing German, I have a german.txt file that ScriptEnter will parse out to translate.
Segue is currently working with Microsoft to come up with a way to recognize window dialogs by control ID which would save a lot of work.
We also approach testing localized (single-byte) versions of our software like TonyM. The only difference is that instead of using a text file to store the strings, we store them in an include file. This, of course, requires the tester to configure SilkTest to "use" the correct include file for the language being tested.
Using TonyM's technique--defining the tag as a variable--does not allow the recorders to later be used; the most important of which is Record Identifier when debugging a "can't find the object" type of failure.
At least this was the case using Silk 5.0.3--OR has this now changed in SilkTest 5.5 or SilkTest I?
Following on from MattGray, this is the approach I have taken in the past. I created scripts which parsed out all the caption tags in the include files, replaced them with variables and put 'varname = "caption"' into a separate include file. This 'glossary' could then be sent off to the translator for translation. My view is that by using some kind of glossary (either text or include) you benefit by having the tool catch (some of the) translation errors.
Unfortunately I do not have these scripts any more, but I do remember they took a couple of days to write and did pay off (we were able to create glossaries for browser include files - IE, Netscape) in about half an hour.
Terry - Recording will indeed work if you resource the strings. Whether you are running the double byte version of SilkTest or not is irrelevant. If your recording isn't working, it's because you've got more going on than just that.