Two wording issues:
- “we are morally obligated to do so
to the best of our ability” à Should we say “we will make a best effort attempt
to do so”? I’m not sure how “morally obligated”
will be interpreted.
- “. WTP will not apply the above
deprecation policy to non-API code where API alternatives exist.” –
We should be clear on what we’re stating here…if migration is
mechanical, do we assume (i.e. force) adopters to migrate?
From:
wtp-pmc-bounces@xxxxxxxxxxx [mailto:wtp-pmc-bounces@xxxxxxxxxxx] On Behalf Of Lawrence Mandel
Sent: Wednesday, November 30, 2005
7:40 PM
To: WTP PMC communications
(including coordination, announcements, and Group discussions)
Subject: Re: [wtp-pmc] Proposal
for Policy on non-API Code Deprecation
Arthur,
Thanks
for the clarifications, especially point 2. I'm fine with this so long as we're
clear to the WTP team that this is the strategy. I agree that quality should be
one of the top priorities in 1.5 (with an emphasis on solving major problems -
such as Java EE 5 support - early in the cycle!) and I would still like to
encourage teams to work on their API in 1.5 as the sooner we define high
quality API the sooner we can remove the requirement to maintain internal
structure.
Lawrence Mandel
Software Developer
IBM Rational Software
Phone: 905 - 413 - 3814 Fax: 905 - 413 - 4920
lmandel@xxxxxxxxxx
Arthur Ryman/Toronto/IBM@IBMCA
Sent
by: wtp-pmc-bounces@xxxxxxxxxxx
11/29/2005 01:24 PM
Please
respond to
"WTP PMC communications (including coordination, announcements, and
Group discussions)"
|
|
To
|
"WTP PMC communications (including
coordination, announcements, and Group discussions)"
<wtp-pmc@xxxxxxxxxxx>
|
cc
|
|
Subject
|
Re: [wtp-pmc] Proposal for Policy on non-API
Code Deprecation
|
|
Lawrence,
The time period between WTP 1.0 and WTP 1.5 is too short for adopters to react
to many changes. We therefore should evolve WTP in non-breaking ways where
possible. We still have ways to move forward in the next release, e.g.
1. Define API, but preserve the non-API code. For example, the API could wrap
the non-API code.
2. Where preserving non-API code is very expensive, consult the scans to
determine who is impacted, review the proposed change with them, and provide
clear migration instructions.
3. Focus on fixing bugs and defer API work to 2.0. We do have a significant bug
backlog. Quality is also extremely important to adopters.
WTP will not be successful unless its adopters are successful. We'll have more
freedom to develop our API for 2.0.
Arthur Ryman,
IBM Software Group, Rational Division
blog: http://ryman.eclipsedevelopersjournal.com/
phone: +1-905-413-3077, TL 969-3077
assistant: +1-905-413-2411, TL 969-2411
fax: +1-905-413-4920, TL 969-4920
mobile: +1-416-939-5063, text: 4169395063@xxxxxxx
Lawrence Mandel/Toronto/IBM@IBMCA
Sent by: wtp-pmc-bounces@xxxxxxxxxxx
11/29/2005 10:03 AM
Please
respond to
"WTP PMC communications (including coordination, announcements, and
Group discussions)"
|
|
To
|
"WTP PMC communications (including
coordination, announcements, and Group discussions)"
<wtp-pmc@xxxxxxxxxxx>
|
cc
|
|
Subject
|
Re: [wtp-pmc] Proposal for Policy on non-API
Code Deprecation
|
|
With the small amount of API defined in WTP 1.0 I understand the need to preserve
non-API to allow for adoption of the platform. However I'm concerned this
blanket support we are proposing for adopters may hinder WTP development as the
team must maintain the internal structure, which may prove more difficult than
maintaining API, while developing new function or simply evolving the structure
of the code towards releasing API.
Lawrence Mandel
Software Developer
IBM Rational Software
Phone: 905 - 413 - 3814 Fax: 905 - 413 - 4920
lmandel@xxxxxxxxxx
Arthur Ryman/Toronto/IBM@IBMCA
Sent by: wtp-pmc-bounces@xxxxxxxxxxx
11/28/2005 05:27 PM
Please
respond to
"WTP PMC communications (including coordination, announcements, and
Group discussions)"
|
|
To
|
wtp-pmc@xxxxxxxxxxx
|
cc
|
|
Subject
|
[wtp-pmc] Proposal for Policy on non-API Code
Deprecation
|
|
PMC Members,
As I discussed previously, I think we need to adopt a policy that helps WTP
adopters evolve as we roll out our API. Here is an updated proposal that
incorporates feedback from David:
WTP 1.0 will not publish API for all functions that adopters may require.
Adopters therefore may elect to use non-API code. Although we are not
technically obligated to preserve the non-API code in WTP 1.5, we are morally
obligated to do so to the best of our ability. We should gracefully deprecate
the non-API code for one or two releases before completely removing it and
forcing adopters to migrate. Since there is a lot of non-API code, we need to
focus on the portions of it that adopters are actually using. Furthermore, we
need to give our developers tools to determine if they are potentially breaking
any adopter. Avoiding breakage in important, frequently used non-API code will
also be good practice for us in preparation for evolving our API code in a
non-breaking way. Here are the main points:
1. WTP will provide to adopters a simple build tool that can use to scan their
code and generate a report of their usage of WTP code that can be attached to a
bugzilla report. The reports will be purely statistical (usage counts) and not
reveal any details about the vendor's code in order to preserve privacy.
Vendors who volunteerily opt in to this reporting process will benefit from our
efforts to reduce potential breakage in their code as WTP evolves its code.
2. WTP will provide to its developers a build tool that scans our code and
detects breakage in both API and non-API code. Known API breakage will be
considered a "blocker" P1 bug. Breakage in non-API code will be flagged
if there is significant adopter usage of it, in which case an effort will be
made to preserve the non-API code, or provide migration documentation. The
techniques used to evolve API code will be applied to the important non-API
code on a best efforts basis. In cases where breakage is unavoidable, affected
adopters will be notified.
3. When feasible, WTP will deprecate important non-API code for a period of at
least one major release. Adopters will be encouraged to migrate to the new
version between releases.
4. WTP will review the non-API usage reports for cases where adopters are not
using available API code and we will inform adopters that API alternatives
exist and that they should migrate to them. WTP will not apply the above
deprecation policy to non-API code where API alternatives exist.
5. WTP will review the non-API usage reports and use that information to
prioritise the evolution of non-API code to API code. Non-API code that has
high adopter usage will be given the highest priority.
Any more comments? If this is agreeable, we should collect the non-API usage
reports in mid-January.
Arthur Ryman,
IBM Software Group, Rational Division
blog: http://ryman.eclipsedevelopersjournal.com/
phone: +1-905-413-3077, TL 969-3077
assistant: +1-905-413-2411, TL 969-2411
fax: +1-905-413-4920, TL 969-4920
mobile: +1-416-939-5063, text: 4169395063@xxxxxxx_______________________________________________
wtp-pmc mailing list
wtp-pmc@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/wtp-pmc_______________________________________________
wtp-pmc mailing list
wtp-pmc@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/wtp-pmc_______________________________________________
wtp-pmc mailing list
wtp-pmc@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/wtp-pmc