I am much more concerned about the two concerns that Nathan raises as well as 
their side effects. And, on the other hand, I see very little benefit to 
releasing 1.3.

In my opinion for the release to be useful to anyone it needs to meet at least 
three requirements:

1.       Meet basic certification for its feature set

2.       Be interopable with previous releases

3.       Be free of major security and quality issues
Without the above three requirements I am not sure the release would be of use 
to anyone. Anyone doing interim work which didn?t have the above requirements 
could just take a snap of the 1.3-rel branch (or even master) and use that for 
their prototype or other work. They wouldn?t really need it to be labeled 1.3 
as such.
However, declaring a 1.3 release at this point would have negative 
consequences. For one, it would be a dead-on-arrival release, where even the 
features it is supposed to have wouldn?t really work properly. Companies trying 
to evaluate or use any of the new features can start building on incorrect 
foundation, and if/when such feature is fixed it may not work for them. For 
those companies or individuals evaluating IoTivity or OCF might conclude that 
the ecosystem is of low quality and not meant to interop with even previous 
releases and therefore untenable for their products. This would also undercut 
the goal of getting more contributors to the project.
I do hear Uze?s concerns that by having the 1.3 release go past its scheduled 
release some of the resources that were originally meant to work may need to 
peel off to do other projects. However, declaring 1.3 done when it really 
isn?t, is not going to fix that problem. The release will still need to pass 
CTT and get its quality up, and declaring 1.3 done will, in my opinion, 
discourage others from joining the effort. Also it will create future headaches 
where it?ll set the precedent of having dead-on-arrival releases. Companies 
wanting to use IoTivity will need to figure out which releases are real and 
which ones are just half-baked. Figuring out which release will work with which 
other releases and to what degree will be a very difficult question to answer 
and would be a big headache for the ecosystem.
In hindsight, I think it would have been better to cut some of the features 
earlier on. This would have made the CTT gap smaller, and would have prevented 
a bunch of the churn and bugs that came with introducing those features late. 
We do still have that lever to some degree though, where if some feature(s) 
proves to be too burdensome to pass we could disable it for the time being and 
move it to the next release. It will not solve the problem to the core, but 
might reduce the time and burden to some degree.

Thanks,
- Omar

From: iotivity-dev-bounces at lists.iotivity.org 
[mailto:[email protected]] On Behalf Of Heldt-Sheller, 
Nathan
Sent: Monday, May 22, 2017 11:10 AM
To: uzchoi at samsung.com; iotivity-dev at lists.iotivity.org
Subject: Re: [dev] Regarding IoTivity 1.3.0 release objective

Hi Uze,

Thanks for doing such as excellent job tracking all these issues and concerns 
and trying to hold to a successful schedule.

In my opinion, as long as people are clear that 1.3.0 cannot be certified as 
either OIC 1.1 or as OCF 1.0 then it is okay to create the release, as a way to 
mark a relatively stable reference for OCF 1.0 feature preview, for purposes 
such as beginning vendor application integration etcetera.   I do not object to 
creating a release, if it is useful for member companies? we want OCF to be 
adopted and be successful, after all!

However I have two concerns:

1.       This ?non-certifiable? 1.3.0 may be used as a basis for more than it 
is intended, and since it is not tested yet against complete 1.0 CTT, I am 
quite certain it contains issues and gaps with the Specifications and their 
required behavior, even if we haven?t identified them yet.  Are we certain 
vendors will not ship OCF 1.3.0 based products and create interop issues in the 
field?

2.       I also am concerned that with the Amsterdam schedule being so tight, 
we will skip ahead and start building Amsterdam on an incomplete foundation of 
1.3.0, without first completing OCF 1.0 implementation (including all bug fixes 
and un-discovered feature gaps).
That said I think if we commit to driving hard to close all remaining gaps and 
bugs with full OCF 1.0 CTT then I don?t see harm in releasing 1.3.0, again if 
it is useful to member companies to do so.  But in my opinion, if we are not 
very disciplined, I think the above two concerns are likely to cause us trouble 
down the road.
Sincerely,
Nathan

From: iotivity-dev-bounces at lists.iotivity.org<mailto:iotivity-dev-bounces at 
lists.iotivity.org> [mailto:[email protected]] On Behalf 
Of ???
Sent: Monday, May 22, 2017 10:35 AM
To: iotivity-dev at lists.iotivity.org<mailto:iotivity-dev at 
lists.iotivity.org>
Subject: [dev] Regarding IoTivity 1.3.0 release objective


Hi All,

Today I heard that there are lots of concern for 1.3.0 release in a hurry.

But I'd like to share some background for this.

Originally IoTivity has a plan to sync up with spec and certifiction schedule.

However, there OCF 1.0 spec has been delayed and eventually this caused the 
misalignment with CTT final and IoTivity 1.3.0 release.

Reply via email to