Sedna XML DB available to websites via Firefox extension

[Brett Zamir]

Hello Dana,

On 7/20/2010 6:18 PM, Daniela Florescu wrote:
> is the following low-level database API for browsers standard useful ?
>
> http://www.w3.org/TR/IndexedDB/
>

Since it is apparently becoming the standard, with both Microsoft and 
Mozilla behind it, IndexedDB will apparently definitely be useful to web 
developers in the sense that it seems it is to become the only option 
there is as far as standard client-side web databases.

Maybe because of its association with complex protocols like SOAP or 
exclusive familiarity with the cumbersome DOM as a means of traversing 
XML (and lack of universal access across browsers to the convenience of 
E4X, XPath, etc.), it is my impression that XML has gotten a kind of bad 
name among RAD developers (if not also for bad blood between 
implementers favoring HTML and standards committees favoring XHTML), who 
prefer to work with structures like JSON, a format which itself is a 
subset of the JavaScript they are familiar with.

However, I also believe that a focus on JSON or JSON-like structures 
puts those of us interested in document-centric data (e.g., TEI or 
collections of books in XHTML) or those who wish to take advantage of 
reliable and standardized namespacing for that matter too, at a big 
disadvantage. The fact that Firefox has 10+ year old bugs for XML as 
regards to staples like entity files/DTDs (e.g., see 
http://pledgie.com/campaigns/7732 ), or HTML5's lack of support for 
ideas such as allowing HTML to be redefined to work better for polyglot 
XHTML/HTML documents (see http://wiki.whatwg.org/wiki/HTML_vs._XHTML ), 
is perhaps another testament to this lack of support for XML in the web 
community (ironic I would say too, given how the web is build on a 
language similar to XML).

Mozilla has been actively updating their own documentation page on 
IndexedDB (at https://developer.mozilla.org/en/IndexedDB ) with someone 
continuing to fill out the pages in just the past few days, making it 
now even more likely that regular web devs like me can begin using it in 
web pages. So at least there will be something we can use (though I'm 
still holding out hope for XML databases, especially if its avid 
off-line users begin to raise their voices in the web apps crowd).

Although IndexedDB allows, like XML, hierarchical storage of data (and 
by its name, I assume indexing as well!) and is also, like XML 
databases, a fashionable "NoSQL" solution, the fact that it is not XML, 
makes it less than convenient to store XML with it.  (Fans of relational 
databases, however, I think, could be convinced to go with XML which can 
use XHTML tables along with JavaScript if not XQuery which simulates 
familiar relational concepts such as joins).

I guess the options are to either take the extra step of converting XML 
into this JSON-like key-value format (and lose out on access to familiar 
querying/adapting technologies as jQuery/CSS Selectors, XPath, XQuery, 
XSL, DOM, etc.) or to simply store whole XML files as strings and then 
pull them out, as one might do if storing in relational databases--both 
far from ideal I would say, for the one not being XML-ish, not allowing 
access to XML tools, and requiring extra steps for conversion, and the 
other for not giving advantages of optimal indexing.

best wishes,
Brett