First Contact

["COOPER, DARREN K GS-14 USAF AFMC 96 RNCS/RNCDR" <da...@us.af.mil.INVALID>]

Greetings,

This is my first contact with any open source community, so please forgive my ignorance.  I am working with a team, trying to solve a problem for the US Department of Defense (DOD), which I think Daffodil is a candidate solution.  Here is a short synopsis:


1.       A contractor develops a specialized data system that produces a stream of data

2.       The data is described, more or less, in some contractor produced ICD documentation

3.       DOD test engineers write code based on the ICD to ingest the new data stream

Wouldn't it be nice if there existed a standard description of the data that could be produced in step 2 which could be used in a universal decoder that replaces the need for step 3.  From my research, I believe daffodil could be that technology.

I am looking for an insider perspective on the state of the incubation and how close it is to being prime time.  If work needs to be done, could DOD developers partner with this community to remove any road blocks.  Finally, is there anyone or a few individuals who are good representatives of this effort that would be willing to participate in a dialog or Q&A on this subject.

If I am not going about this the right way, please redirect me.

v/r
Kelly Cooper

Re: First Contact

[Steve Lawrence <sl...@apache.org>]

Welcome!

Yes, Daffodil certainly feels like it would be a good solution to your
problem.

Regarding incubation, Daffodil currently follows all the quality
standards and development practices that the Apache Software Foundation
requires and depends on that makes their projects successful. The main
reason Daffodil has not yet graduated from an incubator to an Apache Top
Level Project is because of the lack of diversity of companies that
actively contribute. This is an important requirement for graduation so
that 1) if one company no longer desires to contribute, the project can
still continue, and 2) one company cannot dominate the direction of the
project.

So incubation does not necessarily mean low quality or poorly tested
code. From a code quality perspective, I would argue that Daffodil is
currently ready for prime time. In fact, it has been integrated in
numerous solutions that are currently in production.

That said, exiting incubation and graduating to a top level project is
still an important goal we strive to reach. We are always looking for
new contributors that can help us reach our graduation goal and help us
improve the quality of the Daffodil project. Note that there are many
forms of contributions that can be made beyond just code contributions.
For example, documentation, tutorials, helping with questions, bug
tracker triage, testing, etc. Really, we just look for individuals that
are committed to the success of the project, which does not necessarily
mean making commits to the codebase.

DOD developers are certainly welcome to contribute to Daffodil. And we
are more than happy to do what it takes to help new contributors spin
up, regardless of company affiliation. Daffodil is a complex codebase
and DFDL is a fairly large language, so this can all take some time to
wrap ones head around. I know Mike has done one-on-one code screen
sharing sessions to help with that.

To the last question about dialog/Q&A, Apache generally prefers that
discussions happen on a mailing list. Firstly because it makes any
decisions/activity publicly visible (Apache likes to say "if it didn't
happen on a list, it didn't happen"), but also because it can act as a
reference for other contributors that might have the same questions. So
the mailing list is probably the best place to get started if you have
additional questions. However, we have done and are willing to do screen
sharing or one-on-one work to help quicken the spin up and answer
questions, so if that's something you think would be useful we can
certainly try to find the time for that.

As for getting started, if you're interested in code contributions,
probably the first place to start is with the Code Contributor Workflow:

https://cwiki.apache.org/confluence/display/DAFFODIL/Code+Contributor+Workflow

We have also tagged bugs with a "beginner" label, which are bugs that we
believe should be relatively small and a good place to get your feet wet:

https://s.apache.org/daffodil-beginner

Note that even though these are "beginner" bugs, they still usually
require some mentoring to figure out were to get started, so please
don't hesitate to ask if you're unsure.

- Steve



On 9/29/20 1:47 PM, COOPER, DARREN K GS-14 USAF AFMC 96 RNCS/RNCDR wrote:
> Greetings,
> 
> This is my first contact with any open source community, so please forgive my ignorance.  I am working with a team, trying to solve a problem for the US Department of Defense (DOD), which I think Daffodil is a candidate solution.  Here is a short synopsis:
> 
> 
> 1.       A contractor develops a specialized data system that produces a stream of data
> 
> 2.       The data is described, more or less, in some contractor produced ICD documentation
> 
> 3.       DOD test engineers write code based on the ICD to ingest the new data stream
> 
> Wouldn't it be nice if there existed a standard description of the data that could be produced in step 2 which could be used in a universal decoder that replaces the need for step 3.  From my research, I believe daffodil could be that technology.
> 
> I am looking for an insider perspective on the state of the incubation and how close it is to being prime time.  If work needs to be done, could DOD developers partner with this community to remove any road blocks.  Finally, is there anyone or a few individuals who are good representatives of this effort that would be willing to participate in a dialog or Q&A on this subject.
> 
> If I am not going about this the right way, please redirect me.
> 
> v/r
> Kelly Cooper
> 
>