IDE Minutes 2009 03-17

From MemberWiki

Jump to: navigation, search




  • Jon Ferraiolo, IBM
  • Javier Pedemonte, IBM
  • Adam Peller, IBM


Small turnout today

Decision that Jon, Javier and Adam would talk through the issues raised for today's phone call, then try to get decisions via email.

TOPIC #1: Gadgets TF says allow substitution on referenced content

(The second email says allow substitution on CONTENT SRC= and JAVASCRIPT SRC= but not REQUIRE SRC=)


  • It looks like Gadgets TF discussion wants substitution on CONTENT (whether SRC= or not) and JAVASCRIPT (whether SRC= or not). That's acceptable with us. (Let's see if IDE WG will agree with that via email.)
  • We shouldn't do subtitution on REQUIRE SRC=. First off, in many workflows, a REQUIRE element will get translated directly into one line of HTML, such as LINK element (for type=css) or a SCRIPT element (for type=javascript). Secondly, in the IDE case, REQUIRE SRC= assets can be shared across multiple widgets. (Let's see if IDE WG will agree with that via email.)
  • Observation: IDE spec says SRC= is a required attribute on REQUIRE element. (Need to verify that IDE WG is fully aware of this via email.)

TOPIC #2: Gadgets TF worries about @@, %%, and __

Issue summary

OpenSocial Gadgets has:

  • __MODULE_ID__
  • __UP_propertyname__
  • __MSG_localizationkey__

OpenAjax Widgets has:

  • __WID__
  • @@propertyname@@
  •  %%localizationkey%%

and the following syntax for encoding strings:

  • @@entityencode(propertyname)@@
  • @@escapequotes(propertyname)@@
  •  %%entityencode(localizationkey)%%
  •  %%escapequotes(localizationkey)%%


(1) Should our localization strings be consistent with OpenSocial? i.e., instead of %%localizationkey%% we use __MSG_localizationkey__?

(2) Should we go all the way on substitution features to align with OpenSocial?

  • adopt __MODULE_ID__ instead of __WID__
  • adopt __UP_propertyname__ instead of @@propertyname@@
  • adopt __MSG_localizationkey__ instead of __MSG_localizationkey__

(3) What happens with entityencode and escapequotes? Here are proposals submitted so far:

  • (a) __entityencode(__MSG_foo__)__ and __escapequotes(__MSG_foo__)__ (Jon)
  • (b) __MSG_foo(enc: [quo, ent])__ (Howard)
  • (c) __MSG_at_ENTITY_foo__ (Adam)_______________________________________________


Our recommendation is to match OpenSocial's approaches across the board, unless someone comes up with a compelling reason why we can't do that. Therefore, __MODULE_ID, __MSG_foo__ and __UP_foo__. Propose that to the IDE WG

Regarding entityencode and escapequotes, our favorite was what Adam (tried to) propose in email:

  • __MSG@ENTITY_foo__ // for entityencode(). Also works with __MODULE and __UP
  • __MSG@QUOTE_foo__ // for escapequotes(). Also works with __MODULE and __UP

Propose that to the IDE WG

TOPIC #3: preload and postload

Adam: On LIBRARY element or REQUIRE element? More flexible on REQUIRE element. Suppose I want to include a single JavaScript file but set up some config parameters for the JavaScript file before it gets inserted into the document?

Jon: Based on that argument, it makes sense to me to have PRELOAD and POSTLOAD back on REQUIRE rather than LIBRARY. Let's propose that in email

Adam: What about substitution? Aren't PRELOAD and POSTLOAD potentially shared across multiple widgets? Do you know what __WID__ is? Also, in some scenarios, won't PRELOAD be inserted into document before widget is instantiated?

(Conclusion: Don't want to do macro substitution on PRELOAD and POSTLOAD. Let's propose that in email)

Personal tools