I guess the question becomes, "which is easier to deal with, occasional
weirdness from XML (which you have to go back and trace), or manually
merging office documents using tools like xdocdiff?"
The problem I have encountered with manual merges is that users can be
"less intelligent" than the machine, and accidentally leave out
changes. This is especially true for Excel documents, where users leave
out changes to formulas all the time, and then complain about weird
Excel behavior. As long as you understand the pitfalls of XML merges
(which I totally failed to point out), I think XML is the better option.
XML changes are also "diffable" by web tools like trac. We use trac at
my office, so we use XML.
-- Scott
Doug Hockin wrote:
> He may have chosen a bad example, but his point is still true.
> XML can have unordered collections. And there are programs that
> will output the "same" collection in different orders at different
> times. I've seen this many times. It does make it hard to diff
> this sort of XML file. Don't know whether Word does this.
>
> And then there's the issue of different XML that ends up looking
> the same on the screen. So there's no difference to look at
> but the XML is different.
>
> -- Doug
>
>
>> -----Original Message-----
>> From: Andy Levy [mailto:andy.levy_at_gmail.com]
>> Sent: Wednesday, February 13, 2008 12:55 PM
>> To: users_at_tortoisesvn.tigris.org
>> Subject: Re: Merging is not happening for word documents.
>>
>> On Feb 13, 2008 3:41 PM, Ron Wilson <ronw.mrmx_at_gmail.com> wrote:
>>
>>> On 2/13/08, Scott Grizzard <scott_at_scottgrizzard.com> wrote:
>>>
>>>> If you change your Word settings to save documents in XML
>>>>
>> format, you can
>>
>>>> avoid (or at least mitigate) that problem in the future.
>>>>
>> XML documents are
>>
>>>> just text, and svn can merge those.
>>>>
>>> Maybe. While XML files are just text, the order of the elements and
>>> attributes could change without changing the actual content
>>>
>> contained
>>
>>> with the XML markup. In the case of Word document, the
>>>
>> actual textual
>>
>>> content might not change, but the order of some of the "supporting"
>>> elements, or of attributes.
>>>
>>> Example:
>>>
>>> <a id="now"/><div id="3" class="quotation">Now it the time ...</div>
>>>
>>> is equivalent to:
>>>
>>> <div class="quotation" id="3"><a id="now"/>Now it the time ...</div>
>>>
>>> But the XML file is different. And would result in an SVN conflict.
>>>
>> These two examples are not equivalent. In your first example, <a> and
>> <div> are siblings. In the second, <a> is a child of <div> (create a
>> small XML document for each and load them into something that shows
>> the node tree structure like XMLSpy). VERY different. The end result,
>> if transformed to HTML, functions the same, but the XML structures
>> themselves are NOT equivalent.
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe_at_tortoisesvn.tigris.org
>> For additional commands, e-mail: users-help_at_tortoisesvn.tigris.org
>>
>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe_at_tortoisesvn.tigris.org
> For additional commands, e-mail: users-help_at_tortoisesvn.tigris.org
>
>
Received on 2008-02-14 07:21:12 CET