Hopbot log for 2009-04-27 - Helma IRC channel: #helma on irc.freenode.net

2009-04-27:

[9:16] <hannesw> should we rename "helma/system" to just "system" in ng?
[9:17] <zumbrunn> wasn't "system" talked about as a reserved global for the serverjs standard library?
[9:19] <zumbrunn> I guess helma.system does kind of what the intention for "system" in the securable modules was
[9:20] <zumbrunn> https://wiki.mozilla.org/ServerJS/System
[9:22] <zumbrunn> so, yes... +1 for renaming helma.system to just system
[9:26] <zumbrunn> clearly, the "x" prefix thing in that proposal will not fly
[9:30] <zumbrunn> also, I'm noticing that the votes for system as a module were not counted properly
[9:30] <zumbrunn> system as a module won that vote, I believe
[10:14] <hannesw> yes, I'll enhance system with the proposed properties i think
[18:25] <hannesw> I'm about to rewrite the texts for the ng demo app in markdown
[18:26] <hannesw> some help with improving the texts would be great.
[18:26] <hannesw> http://helma-ng.appspot.com/
[19:05] <liamMT> hannesw: what kinda improvements did you have in mind?
[19:05] <hannesw> liamMT: Improving the text: provide a better introduction, better explanation of features, etc.
[19:06] <liamMT> I'll see if any spring to mind :)
[19:53] <hannesw> thanks!
[22:41] <liamMT> hannesw: for improved text, one helpful thing could be an overview on the front page.
[22:41] <liamMT> this is missing from the ng wiki front page as well
[22:41] <hannesw> liamMT: yes, i agree
[22:41] <liamMT> and the beginning of the ReadMe in the source distribution isn't too bad
[22:41] <liamMT> might work for that
[22:42] <hannesw> yes, i rewrote that recently. may be a good starting point
[22:43] <liamMT> obviously the profiler page needs a bit more, but I was wondering - do you have a sense of the performance difference when running rhino in fully interpreted mode, rather than compilation mode?
[22:44] <liamMT> is there some general proportion/scaling factor that might be useful, or is it too tough to say? because being able to say, "run this in the profiler and you can trust that your compiled version is pretty much the same, just 25% faster" or something like that
[22:44] <hannesw> i can't put any numbers on it. i think it depends a lot on the code you're running
[22:45] <liamMT> yeah, I figured
[22:47] <hannesw> yes, the profiler is going to be slower, not only because it's in interpreted mode, but also because of the profiling itself, which is mostly implemented in JS
[22:47] <hannesw> but I don't think that's a problem. What you get is the relative hot spots of your code
[22:53] <liamMT> right - I guess the urge to use a profiler as a benchmarker is not quite right

 

 

In the channel now:

Logs by date: