XQuery and immortality WAS ACID transactions WAS Re: from Where 2.0

[Daniela Florescu]

>  Michael R is saying, I think, that
> transaction facilities (and scripting) should only be added to the  
> standards
> once they have been proven in products.

Michael (K),

thanks for the translation ! :-)


Michael (R),

I am sorry, despite the perfectly understandable English, I don't  
understand the logic.

I read, and reread the sentence above, and despite the rewriting, the  
content still doesn't makes sense to me.

Let's take scripting from your email for example:

- We have 30 years of experience with scripting extensions of  
declarative languages
(PL-SQL, TrnasctSQL), and millions of customers across the globe.

- in functional languages, we've beatten down how to extend "pure"  
functional languages
with side effects, for 20 years at least (the list of such languages  
is endless)

- in the XML land, XQuery scripting extensions have been studied in  
research since
XQuery was Quilt (2000)

  - spec-ed and implemented first time in 2001, used in production by  
large customers at that time

- shipped by companies (BEA for example) since 2004, used widely, again

-  ";"  implemented in MarkLogic again since the beginning of time and  
widely used

- multiple standard submissions to W3C since 2006

- lots of discussion inside the W3C WG

- working drafts since 2007

- W3C working draft implemented in at least 3 engines that i know, and  
used by customers in real
world

- active research being done in many places in compilation and  
execution since 2005 (Sigmod papers, etc)

- we are mid 2010

====


After all that, we are not "proven" yet. Sorry, I fail to understand  
what else is needed to be
labeled "proven" !? And why we should wait another  10 years for  
simple ";" and a variable assignment.
  What is there not known yet by any computer scientist !?


=====

I would start with another list about how many think we know (or  
should know) about transactions
and consistency,  but I am afraid it gets longer, and I need to run to  
the office.

I personally run our of time before retirement. I simply cannot afford  
to wait another 15 years before we add a
commit() or checkout()/checkin() to XQuery.

That would explain my impatience, I guess :-)

Best regards
Dana