W3C home > Mailing lists > Public > public-appformats@w3.org > November 2007

Re: [Widgets] widget resources

From: Jon Ferraiolo <jferrai@us.ibm.com>
Date: Wed, 28 Nov 2007 16:57:24 -0800
To: "Marcos Caceres" <marcosscaceres@gmail.com>
Cc: "Bjoern Hoehrmann" <derhoermi@gmx.net>, "public-appformats@w3.org" <public-appformats@w3.org>, public-appformats-request@w3.org, "Thomas Roessler" <tlr@w3.org>
Message-ID: <OFF18042CF.F0DB2924-ON882573A2.0004818A-882573A2.00054156@us.ibm.com>

Hi Marcos,
I skimmed the new text and so far I have only seen two things that makes me

(1) Why does the note about path lengths talk about 100 characters rather
than 255 characters?

(2) Why are we telling people they must not have "CON" or "PRN" (etc) as a
folder or basename? Because this causes problems on one of today's
operating systems? If that's the reason, then I disagree with this.
Instead, don't legislate away any particular names, and instead put in
another note that talks potential interoperability on some systems, similar
to the note about 100 character paths. Similarly, does the ZIP appnote
preclude leading or trailing spaces, or is this just a problem with
particular OS's, in which case this should also become a note instead of a
conformance requirement. Same question about all of the "reserved
characters" (<, >, :, etc)


             "Marcos Caceres"                                              
             mail.com>                                                  To 
             Sent by:                  Jon Ferraiolo/Menlo Park/IBM@IBMUS, 
             public-appformats         "Thomas Roessler" <tlr@w3.org>,     
             -request@w3.org           "Bjoern Hoehrmann"                  
             11/27/2007 02:39          <public-appformats@w3.org>          
             AM                                                         cc 
                                       [Widgets] widget resources          

Hi All,
Taking into consideration Bjoern's, Thomas', and Jon's feedback, I've
re-written the Widget Resources section of the [1].

Bjoern, I've trimmed much of the text. Please let me know if that is
any better?

Thomas, I made sure the ABNF now directly follows to the definition of
a zip relative path. I've also tried to address the ASCII and UTF-8
issues by being quite restrictive. From my own testing, both Zip and
Widget implementations seem pretty messed up when it comes to
processing file names outside ASCII... so what I've spec'd was the
best solution I could come up with: if it says it's ASCII, but it's
not, then die; otherwise, feel free to treat as UTF-8. However, Art is
setting up a discussion with the internationalization folks about this
issue to see if they can come up with a better solution.

Jon, all instances of the word "preferred" have been replaced to
"recommended" throughout the document. The text about path lengths has
now become an informative note.

I'd really appreciate if you could take the time to read the updated
text and let me know if it is ok (please ignore any markup/style
errors, I'll fix those soon).

Kind regards,

[1] http://dev.w3.org/2006/waf/widgets/#widget0
Marcos Caceres

(image/gif attachment: graycol.gif)

(image/gif attachment: pic20361.gif)

(image/gif attachment: ecblank.gif)

Received on Thursday, 29 November 2007 00:59:18 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:50:08 UTC