taverna-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dmitry <redmi...@list.ru>
Subject Re: Common workflow language
Date Fri, 04 Dec 2015 16:26:43 GMT

I am quite skeptical about any new workflow language given that we 
already have something like 50...
How one would describe XML message in ontology (not saying it is 
What kind of services workflows should support?

I may be wrong, but CWL looks like a command line pipeline oriented 
script processor rather than a "language".

If one is looking for something new to be used as a "workflow language" 
I would rather take a look at the jBatch (jsr352) specification.
There is still a time to join a JCP for the next jBatch spec (IMO they 
need to separate input/output properties).

It is XML / Java based, but the XML part is quite abstract to be 
implemented in any other language.

The idea behind is that there are several components

Abstract API - Java interfaces that must be implemented by "processors"
Simple XML language that defines the workflow in terms of jobs, batches, 
The configuration that defines aliases linking workflow blocks to 
concrete implementation.
And, finally, the "runtime" that executes the workflow (part of JEE7 or 
standalone). Note that one could develop its own runtime to execute in a 
cloud, grid, etc.

Note that in case of some step failure, it is possible to restart the 
step. There is also a way to specify steps that could be done in parallel.

Using jBatch it is possible to define a set of processors - 
"wsdl_executor", "rest_executor" that then can be used by the workflow.

I played with it to execute a simple workflow that link two web 
services, but found some limitations - no separation between 
input/output parameters and
no parameter types.



On 12/4/2015 4:28 PM, Ian Dunlop wrote:
> Hash: SHA256
> Hello,
> Does anyone have views on how Apache Taverna fits into the goals of
> the Common Workflow Language project
> https://github.com/common-workflow-language/common-workflow-language ?
> Should taverna adopt CWL instead of SCUFL2. Are there things that
> SCUFL2 or the current taverna engine offers that CWL does not?
> Cheers,
> Ian
> - -- 
> Ian Dunlop, eScience Lab
> School of Computer Science
> The University of Manchester
> http://orcid.org/0000-0001-7066-3350
> Version: GnuPG v2
> fCN+KAtD4UC0l2rtynXsXSzr98Vp8/iisuLNUXAw+BO0Xc4oXBmU6QtgDfYnv/tB
> uf6MQJsoE/zvFVjslY3vKWw3IJS7iyIWuGl4caG/lUEUdiQFRRq57e/WFM+IoPqG
> jE6UO3JUmZJw/GVBJqkkoqs4p5EdTArGWz3/qICkl1UkUEEqgtHjlJO9APl5Szfv
> 9X6n1SV5BcmH0XXd1YIHBW+s6pSCKbxRjOhPURSN+FTmMYSM0kXzeD9c2MnzhGU=
> =Agk2

View raw message