Re: meeting followup: (1) Monetate is hiring, (2) notes on jslint v. closure

[Dan Mead <d.w.mead@gmail.com>]

also, for message validation i recommend using a combinator-parser
library instead of an xml schema

http://jparsec.codehaus.org/

This lets you define a parser which looks like a BNF specification for
your language (you're treating the message format as a language)
this is really good when you want to read in a language to an AST to
do some back end manipulation, but you can also just use it as a
filter for correctly formed messages.


*i've never used jparsec. but i have used parsec for a bunch of projects

-Dan

On Fri, May 28, 2010 at 10:34 AM, Hunter Blanks <hblanks@monetate.com> wrote:
> All,
> To follow up from two remarks I made at the meeting:
> (1) Monetate is looking for both senior and junior people to join its
> engineering team. If you want to work with great people -- or you know
> someone who might -- please take a look at:
> https://www.ventureloop.com/ventureloop/jobdetail.php?jobid=38828
> (2) Regarding jslint and Google's closure compiler -- Kyle is correct, and I
> was wrong: both jslint (when given undef: true) and the closure compiler
> will check for undeclared variables. In practice, we use both of them at
> Monetate. Jslint alone checks for stylistic issues like missing semicolons,
> whereas the closure compiler does a better job of following code paths, can
> type check function arguments, and can remove unused functions from the
> compiled output. This last feature (as well as closure
> library's calcdeps.py) make it a lot easier to write modular JS libraries
> and bind them in as needed at compile time.
> Thanks for reading,
> Hunter
> --
> Hunter Blanks
> Monetate, Inc.
>