[uf-discuss] Live Clipboard and uF escaping (was Fwd: 0.91
Speccomment: escaped markup is harmful)
connect at stevenR2.com
Thu Apr 6 06:09:29 PDT 2006
I have only indirectly picked up on this but...
If we consider the web/data as is just now, then the serialization aspect perhaps makes some sense, although the Xml/microformat structure is obviously preferible.
But with pure DOM access it may be tricky to populate an XHTML or other structure as you don't really have a clue as to whathe DOM should do - uless you assumed that everyone who has some non-schema based structure formatted their html in the same way.
In this scenario, at least a serialized version could be used to say "this is a serialization of type X" - very hard to determine i'd say from pure DOM access. This i guess would be useful for people who have existing common formats (some of which you may not wish to publish directly to sa web page - vCard format directly placed in a section of a web page is just nuts - i just tried) who wish to publish, but who don't want to transform all to a Microformat (yet).
I remember the days when everyone was going to publish their PO's in a single Xml Schema format.
Then again, i may have missed the point :)
---------- Original Message ----------------------------------
From: "Danny Ayers" <danny.ayers at gmail.com>
Reply-To: Microformats Discuss <microformats-discuss at microformats.org>
Date: Thu, 6 Apr 2006 12:25:59 +0200
>The post below is from the MS Live Clipboard list,
>in relation to:
>I'm not sure I understand why Matt suggests XML data might have to be
>delivered as both XML and escaped as well, but he gets into
>browser/DOM territory, a place presumably well-known around this list
>- thoughts appreciated.
>---------- Forwarded message ----------
>From: Matt Augustine <matta at microsoft.com>
>Date: Apr 6, 2006 5:23 AM
>Subject: Re: 0.91 Spec comment: escaped markup is harmful
>To: LIVE-CLIP at discuss.microsoft.com
>Thanks for all the information. Escaping only non-XML data formats and
>leaving XML data formats as part of the Live Clipboard document seems
>like a reasonable compromise, but I have a few reservations:
>Since the items in the LiveClipboardContent object might contain either
>XML or escaped, non-XML data, we would most likely have to add a second
>property, named something like xmlData, to hold an XML node in addition
>to the existing data property. Applications would have to know which
>property to use based on the contenttype of the format.
>In the microformat case, treating the data as a string rather than as
>parsed XML makes it easy to display the data by setting it as the value
>of the innerHTML property of an element on the page. In order to avoid
>forcing applications that use this technique to first serialize the
>xmlData, we would probably want to always provide this serialized data
>in the existing item data property. If there is a way to directly add
>the parsed XHTML of the microformat into an HTML element we could avoid
>this, but as far as I know this isn't possible.
>IE6 provides an easy way to get the serialized string value of all the
>xml content of a particular node via the "xml" property. I'm not aware
>of a similar property when using the DOMParser in Mozilla or other
>browsers. Is there an easy way to get access to this serialized data in
>order to satisfy the previous requirement?
>As for David's comments about the ugly XML parsing / serialization code,
>we tried to contain all of that error-prone stuff within our script to
>make interfacing with Live Clipboard data as simple as possible for the
>developer and to make sure the emitted XML put on the clipboard is
>compliant with the spec. We're certainly open to alternative light
>weight implementations of the control though, as long as they are
>compatible with the clipboard XML format as specified and present the
>same user experience for copy/paste.
>Software Design Engineer
>CTO Concept Development Team
>From: Live Clipboard discussion and feedback
>[mailto:LIVE-CLIP at DISCUSS.MICROSOFT.COM] On Behalf Of M. David Peterson
>Sent: Tuesday, April 04, 2006 7:41 PM
>To: LIVE-CLIP at DISCUSS.MICROSOFT.COM
>Subject: Re: 0.91 Spec comment: escaped markup is harmful
>On Tue, 04 Apr 2006 16:24:51 -0600, Danny Ayers <danny.ayers at GMAIL.COM>
>> On 4/4/06, Seth Russell <russell.seth at gmail.com> wrote:
>>> In many applications the programmer is not in control of what kind of
>>> will be processed through the system.
>> But to be able to process it at all, some indication of what it is
>> will be needed - and the current spec has the contenttype attribute.
>> If it's an XML format, it should be indicated here by the producer.
>> This is independent of whether the data appears directly as XML or is
>Exactly. To be able to convert it from one format to another, an
>important piece in all of this is knowing what type of format you are
>dealing with in the first place. To think that the oddball case in
>we have absolutely no clue whats on the clipboard should be the driving
>force behind using a politically-"I know no type I do not like"-correct
>processor solution is quite simply (and being quite frank) -- stupid.
>Don't mean to offend, just don't want to spend the next 10 years of my
>life dealing with code that makes me itch just thinking about it when I
>know there are a VAST number of better solutions than can and should
>be/have been implemented.
>> I don't think that there is a way to
>>> always convert content from the wild it into well formed XML. If
>>> please point me to a easy to deploy PHP library.
>> What source format? HTML Tidy can output XHTML, I'm sure there's a
>> binding for PHP. I don't know of any pure-PHP options, but I'd be
>> surprised if there weren't any.
>Yep. In fact there is both a tidy and, if not mistaken (though need to
>check to be certain) a Tag Soup solution for PHP. In fact, being one
>has been working quite intimately and directly with both Saxon and Dr.
>to help bring the Saxon.NET project over to Saxonica, to then work on
>various details with the new implementation, I can tell you with solid
>assurance that there is a simple and straight forward way of invoking a
>Tag Soup transformation of any given file, or folder of files/file
>folders, with one extension function built directly into the Saxon
>engine. Given the ability to invoke Java-based solutions and, with
>Phalanger, .NET-based solutions from within your PHP code base, then it
>can easily be stated that "yeah, there's a solution that will work with
>> If Live Clipboard is to
>>> be useful between domains in the wild, i think the CDATA "black hole"
>>> is the
>>> only way to go.
>> I'm not suggesting that every producer should produce XML. Only that
>> *when they do* it's transported as such and not obfuscated.
>Yep. What Danny said :)
>Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
>microformats-discuss mailing list
>microformats-discuss at microformats.org
More information about the microformats-discuss