On Tue, 22 Nov 2005, Jason R Huggins wrote: [...] > For this kind of scenario, the table style of Selenium is limiting. I'm > working on a new syntax for Selenium tests... basically a > re-implementation of HyperTalk. Table tests would still be supported for > the simple stuff. This new syntax would be an addition to the family. > > I call this new language "SweetTalk". For your example, the code would > look like the following: [...]
How is this an improvement on the earlier idea of adding a little support for writing tests in JavaScript? I imagine such support would be very simple to implenent (perhaps it's already there and only needs documenting, or simply a published example or two). My thoughts run like this: - Lots of people know JS already. - JS is easy to learn. - JS itself (as opposed to the unpleasant, poorly standardised libraries and built-in data types browsers come with) is quite a nice general-purpose language. - What, specifically, are the advantages over offering a bunch of ten or so JS functions, or a few classes (OK, prototypes not classes -- whatever ;-)? - I have a hard time buying the argument that a general-purpose, well-designed language is not useful for testing :-) Your language doesn't seem aimed at being such a thing (I hope not, anyway!-). - Even it I'm wrong on the previous point (I know others will disagree with me), I often suspect that the learning-time cost of these little languages is often under-reported. - Let's get more great free stuff out of Jason by diverting his time into something more useful ;-) the-first-thing-we-do-let's-kill-all-the-little-languages-<wink>-ly y'rs, John _______________________________________________ Selenium-users mailing list Selenium-users@lists.public.thoughtworks.org http://lists.public.thoughtworks.org/mailman/listinfo/selenium-users
