Good, this is what I wanted to know when I originally sent the email out. Now
its more clear.
I laid out my roadmap and im going to be completely focused on it this year.
Remember, this is a data first project. Its in our charter.
I am going to ask that you work with me and move forward as best as possible.
This project is too young and small to be able to properly deal with so many
competing goals. If you are unhappy with this, then I would ask that you step
down from the PMC. Its going to be a lot easier for me to steward this project
without your distractions.
Obviously this project is a democratic process and if you feel its in the best
interest to keep completing APIs, then thats going to be upto you and the PMC.
I just ask that you stop promoting and marketing one over the other. I only get
involved in these discussions in response to your statements. Otherwise, I
moved forward from the legacy client back in 2011 when I first engaged with
OpenDDR project and wrote dClass. I really havent really looked back since. The
fact that its 2015 and you are promoting the legacy client as a better
alternative is just a bit concerning, thats all.
From: Werner Keil <[email protected]>
To: [email protected]; Reza Naghibi <[email protected]>
Sent: Thursday, January 1, 2015 1:12 PM
Subject: Re: https://issues.apache.org/jira/browse/DMAP-112
Since the DDR Simple client was contributed and accepted by Apache as initial
part of DeviceMap there is no useful path to release it under OpenDDR.org.A
majority of issues were related to e.g. the domain, well that's been abandoned.
And see Volkan's request, the "new" DeviceMap Java client just like everything
else (except data 1.x) was NEVER properly released under Maven either,
Having a "new" client incompatible and broken under Java 8 (while the "legacy"
initial donation which should have been released 2 years ago works perfectly
fine under Java 8 or even 9;-) is a very big concern and does not make it look
good or mature until such bugs are resolved.
Everything that matters to the W3C DDR implementation is in Apache JIRA now,
too, so any GitHub ticket or old site from 2012 are pretty much irrelevant.
Bertrand/all,
According tohttp://httpd.apache.org/dev/guidelines.html
Release Plans (also what Reza sketched, that isn't more than an "intent" since
there has been no vote on it so far;-) require Lazy majority (at least 3 x +1
and more +1 than -1) decides each issue in the release plan
It's a bit hard to say for a very small PMC like this one, so what are your
intentions in this case?
The W3C DDR Simple client has and always had some advantages the "new" client
may never get or at most in a 2.x release some time in 2016 or beyond.-
Recognition of close enough devices based on the "confidence level", so while
several of the tickets and UAs completely unrecognized by Classifier or at most
saying "Generic Samsung Device" (with a 320x250 screen or so;-O) DeviceMap W3C
Simple DDR will return the best compatible device from such family, e.g. a
known "Samsung T530" can be identified with important aspects like screen size,
etc. based on existing "T520" or "T520" providing a better result than
Classifier for screen layout, etc.- True OS Version thanks to a more
specialized Builder system, where a factory default OS is outdated, it'll be
substituted with the actual OS of this device.- Works on Java 8 without bugs or
glitches as of now!
Some of these won't be "rewritten" by Reza till Java 9 or even 10 comes out, so
why defraud the community of a client that offers all that now and did so ever
since early 2013 in fact.
A negative vote on DeviceMap Simple DDR 1.0.0 would mean to gain the community
to access these and not lose them (forever or a very long time) it had to be
removed from DeviceMap and launched as OpenDDR Java Client 1.0.0.29 or rather
1.1.0 (or even 2.0 who cares;-) with dependency to Apache DeviceMap Data.
Unless of course a majority here preferred all of device-data to be TAKEN DOWN
immediately and moved to OpenDDR-data while the DeviceMap users will have to
wait for a working 2,0 rewrite of data, too?;-O
DeviceMap would immediately lose the quality stamp "W3C Compatible" because
just having some "legacy crap" XML that was inspired by W3C DDR without
implementing the W3C Simple DDR API does not mean any of that will ever be W3C
compliant. It simply isn't, leaving that quality stamp to OpenDDR .next.
WDYT?
Werner
On Thu, Jan 1, 2015 at 6:34 PM, Reza Naghibi <[email protected]>
wrote:
>> Everything that's not under "contrib" should have at least one official
>>release.
Just to be clear, im going to be voting -1 on any legacy client release. The
legacy client can be downloaded and supported here:
https://github.com/OpenDDRdotORG/OpenDDR-Java
The above project can be the branch you are talking about. I dont see any value
or upside in re-releasing another legacy client branch under DeviceMap. Have
any of the issues [0] been resolved? What about the pull requests [1]? Lots of
reasons to move on from this client. Then we have the fact that 2.0 will render
it obsolete.
[0] https://github.com/OpenDDRdotORG/OpenDDR-Java/issues
[1] https://github.com/OpenDDRdotORG/OpenDDR-Java/pulls
From: Werner Keil <[email protected]>
To: [email protected]; Reza Naghibi <[email protected]>
Sent: Thursday, January 1, 2015 12:25 PM
Subject: Re: https://issues.apache.org/jira/browse/DMAP-112
Everything that's not under "contrib" should have at least one official
release. That's why I suggested to move the "test-data" folder into /contrib
(unless we agree it shall better be deleted from /trunk?)
On Thu, Jan 1, 2015 at 6:23 PM, Werner Keil <[email protected]> wrote:
OK thanks.
DDR Simple 1.0.0 is ready for release pretty much as of now. A few minor
improvements include using only Log4J 2 instead of SLF4J ("Eat your own
Dogfood" at Apache if you want, all examples are already migrated;-) The only
dependency is Data 1.0.1 or above, aside from the embedded W3C JAR.
A combined Examples project 1.0.0 shall follow. Demonstrating all Java
libraries.
Ideally we should have Jenkins CI sessions for 1.x and later 2.x branches of
all relevant artifacts, at least "Java" and "Examples". The Java 8 glitch
showed, we must have more than just Java 6 even if this is the minimal JVM
version we support until further notice;-)
Werner
On Thu, Jan 1, 2015 at 6:16 PM, Reza Naghibi <[email protected]>
wrote:
Ok, just making sure.
So here is my rough roadmap:
-Jan 2015 - release data 1.0.2-Q2 2015, release data 1.0.3-Q3 to Q4, release
data and client 2.0
From: Werner Keil <[email protected]>
To: [email protected]; Reza Naghibi <[email protected]>
Sent: Thursday, January 1, 2015 12:05 PM
Subject: Re: https://issues.apache.org/jira/browse/DMAP-112
That's why it is slated for a 1.0 release (regardless of maturity but it goes
along with 1.x of the data files) and it does not need to be compatible with
any 2.0 data changes in future.Whether or not we maintain 2 branches of data,
guess time shall tell. Patches or fixes that can be applied maybe, but like
with other vendors (DeviceAtlas, 52DegreesMobi or ScienteMobile/WURFL) they may
not maintain 2 parallel repositories forever.
Werner
On Thu, Jan 1, 2015 at 6:02 PM, Reza Naghibi <[email protected]>
wrote:
I understand.
What is the roadmap for the legacy client?
Just so you are aware, 2.0 data release will remove the notion of a builder
from the data. How will the legacy client evolve to support this? Take a look
at this:
http://svn.apache.org/viewvc/devicemap/trunk/devicemap/java/simpleddr/src/main/java/org/apache/devicemap/simpleddr/builder/?pathrev=1648789
Go into the os, browser, and device. There is no way 2.0 data can maintain
support for this kind of builder architecture. Are you aware of this?
From: Werner Keil <[email protected]>
To: [email protected]; Reza Naghibi <[email protected]>
Sent: Thursday, January 1, 2015 11:48 AM
Subject: Re: https://issues.apache.org/jira/browse/DMAP-112
Well, it's just a fact, the W3C Simple implementation shall be released as 1.0
according to a discussion we had a long time ago.
It's been around and stable almost ever since OpenDDR donated it and that has
not changed.The only reason it was not released was an incompatible change
(bug) you slipped into a Data release 1.x where packages of builders were named
improperly.This was fixed with 1.0,1, thus nothing prevents a release of the
DDR artifact.
There is no rift, we have a .NET client (which is barely maintained, but that's
another issue) and a new direction which allows more flexibility in the
"Cloud", e.g. retrrieving device data from various sources. That's something
W3C did not consider important, so the location of the data source has to be
fixed, that's the only real drawback of the W3C DDR implementation. It could
only be fixed in the W3C code we don't own;-)
I trust the PMC is wise enough to release the W3C DDR implementation, because
not only examples depend on it. And it provides functionality and recognition
power the "New Cloud" client will not get until V 2.0 as you also said. It does
not matter if that happens in JIRA or the mailing list, JIRA is also
transparent and allows us to see what's been done;)
Cheers,Werner
On Thu, Jan 1, 2015 at 5:39 PM, Reza Naghibi <[email protected]>
wrote:
So I don't want to have discussions in JIRA, so Werner, can you please
elaborate on your last point in this thread:
"Indicates, while not in direct competition, the "older" W3C one is a bit more
mature and stable."
https://issues.apache.org/jira/browse/DMAP-112
The reason im asking you to elaborate is because as a PMC member, I would like
to better understand your thoughts on this project, its mission statement, and
where we are going with our data and API development.
Im concerned because you seem to be building a rift in this project in regard
to old API vs new API. This is not going to be good for the health of the
project so this is an attempt to better understand your views and feeling
towards the project and the direction we are going in. This is important
because you are someone who actively promotes and markets this project.
