[DAS2] DAS vs. Java Web Services

[Nomi Harris]

On 10 January 2006, Thomas Down wrote:
 > On 9 Jan 2006, at 22:14, Nomi Harris wrote:
 > 
 > > i was talking to someone in the apollo community about writing a DAS/2
 > > adapter for apollo, and he said:
 > >
 > >> I think the Java Web Services model has been proven more so than  
 > >> the DAS
 > >> model for standardized web application communication.
 > >
 > > this is an interesting claim, and i was wondering if you have any  
 > > comment
 > > on why this is wrong (i assume you all disagree with it, or else you
 > > wouldn't be working on DAS/2).  i think we should address this  
 > > issue in
 > > the grant--maybe we can add it to the progress report?  otherwise, the
 > > review committee could come back and say, "why don't you just use Java
 > > Web Services instead of developing this new standard?"
 > 
 > It would be interesting to know a little more about what's mean by  
 > "Java Web Services".  I'm guessing that you contact is talking about  
 > "web services based on SOAP, WSDL and associated technologies".   
 > There's nothing Java-centric here, although I guess it's true that a  
 > lot of the advocates for these technologies are coming from a Java  
 > or .NET background.
 > 
 > DAS (and DAS/2 in particular) aims for a RESTful style of interaction:
 > 
 >                http://www.ics.uci.edu/~fielding/pubs/dissertation/top.htm
 > 
 > I think the main argument for this was that REST services, which put  
 > all the query information in a URL, are easier for "casual"  
 > programmers (who aren't familiar with SOAP client libraries and the  
 > like) to script about -- and it's even possible to fire a query off  
 > by typing a URL in your web browser.
 > 
 > The downsides of REST are that you get less benefit from existing  
 > client libraries and server frameworks, and that (at least in my  
 > opinion) it's harder to clearly express a rich query language as URLs  
 > than as POSTed XML fragments.
 > 
 >              Thomas.

i certainly agree that "java web services" is vague and it's not clear
exactly what he meant by that.  what i got from his statement was a
knee-jerk reaction against DAS, and i think that's something we need to
take seriously, because the people who review our grant could also think,
with or without justification, "why are they working on DAS when web
services is a better way to go?"

i see at least three aspects of the DAS vs. web services debate:
1. which technology is easier for programmers to use?
2. which delivers the data faster (and/or more flexibly)?
3. what is the public perception of the two technologies and the
availability of servers?  even if DAS is easier to program AND faster, if
most labs are choosing to make their data available via web services (or
some other DAS alternative), we're going to be fighting an uphill battle.

i think it is worthwhile for us to ponder these issues for our own
clarity of purpose, but also so that we can convincingly argue to the
grant reviewers that they should be supporting DAS/2.

      Nomi