2006 October Members Meeting Minutes

From MemberWiki

Jump to: navigation, search

The following are the minutes from the two-day OpenAjax Alliance face-to-face meeting that occurred at Sun's Santa Clara campus on Oct 5-6, 2006.


Contents

Attendees

  • Adam Peller <apeller@us.ibm.com>
  • Alex Russell <alex@dojotoolkit.org> (Dojo)
  • Andre Charland <andre.charland@nitobi.com>
  • Andrei Dragomir <adragomi@adobe.com>
  • Anil Sharma <anil@vertexlogic.com>
  • Chris Erickson <chris.erickson@icesoft.com>
  • Chris Matthews <cmatthews@elinkbiz.com>
  • Chris Schalk <chris.schalk@oracle.com>
  • Christophe Jolif <cjolif@ilog.fr>
  • Coach Wei <coach@nexaweb.com>
  • Craig McClanahan <craig.mcclanahan@sun.com>
  • Daniel Nguyen <dan.nguyen@openwave.com>
  • David Boloker <boloker@us.ibm.com>
  • David Stanek <DStanek@ag.com> (American Greetings)
  • Erwan Paccard <epaccard@ilog.fr>
  • Florian von Kurnatowski <Florian.von.Kurnatowski@scalix.com>
  • Greg Wilkins <gregw@webtide.com>
  • Håkon Wium Lie <howcome@opera.com>
  • Ian Wenig <ian@zoho.com>
  • Jeremy Chone <jchone@adobe.com>
  • Jochen Krause <jkrause@innoopract.de>
  • John Crupi <john.crupi@jackbe.com>
  • John Janetos <jjanetos@laszlosystems.com>
  • Jon Ferraiolo <jferrai@us.ibm.com>
  • Joonas Lehtinen <joonas.lehtinen@itmill.com>
  • Joseph Becker <bjoseph@us.ibm.com>
  • Kevin Hakman <khakman@tibco.com>
  • Lindsey Simon <lsimon@finetooth.com>
  • Mike Milinkovich <mike.milinkovich@eclipse.org>
  • Mike Pinette <mikep@zend.com>
  • Raju Vegesna <raju@zoho.com>
  • Sam Silberman <sam.silberman@openwave.com>
  • Scott Dietzen <scott@zimbra.com>
  • Shane McMillan <shane.mcmillan@softwareagusa.com> (SoftwareAG)
  • Stephen Maryka <steve.maryka@icesoft.com>

Governance and IPR Policy

The meeting was conducted under the terms of the Members Agreement. Only representatives of companies that had signed the Members Agreements were allowed to attend the meeting.


Agenda

The agenda for the meeting can be found at: 2006 October Members Meeting Agenda


Minutes

Welcome Session

  • 9am-10:30a
    • Welcome and Opening Remarks ("We have come a long way baby!")
      • David Boloker (IBM) and Scott Dietzen (Zimbra), two founders of the OpenAjax initiative
    • Announcement of New Steering Committee
      • The first (non-interim) Steering Committee will be announced by the Interim Steering Committee: David Boloker (IBM), Mike Milinkovich (Eclipse), John Janetos (Laszlo), John Lilly (Mozilla), Mike Pinette (Zend), Scott Dietzen (Zimbra)
    • Accomplishments since May
      • Alliance operations - Jon Ferraiolo (IBM)
      • Marketing committee - Jon Ferraiolo (IBM) [unless someone else volunteers]
      • Technical committees - Coach Wei (Nexaweb)

Opening remarks

Dietzen: preserve web 1.0 value proposition for ajax: any server, any browser, any os.

Boloker: ajax conf yesterday - negative press on ajax regarding security. 220+ oss projects, 20-50 commercial oss apps. same issues as j2ee in 1998 -- reliability, scalability, servicability, security. everyone wants to move ajax forward. google/yahoo have proven scalability. reliability issues persisist on the web. security needs a look. avoid web services approach. next steps - how do we interoperate? what do we do about security? about mobile ajax?

Dietzen: getting our apps to work isn't enough -- we need to reach the vb/powerbuilder/flash developer community in ease of use and tooling.

Vote results

Jon Ferraiolo: 11 nominees: Dojo, IBM, Eclipse, ILOG, Laszlo, Nexaweb, Openwave, SAP, TIBCO, Zend, Zimbra

27 votes counted / 53 member organizations

3 votes late or not signed agreement

2 year terms: DOJO (Alex Russell) Eclipse (Mike Milinkovich) IBM (David Boloker) Zimbra (Scott Dietzen)

1 year terms: NEXAWEB (Coach Wei) TIBCO (Kevin Hakman) ZEND (Michael Pinette)

recursive algorithm chose winners.

Q: elections every year? A: 1- and 2-year terms. Next year, expiring slots will be up for 2 year terms, and only 2 year terms going forward.

Q: More balance between open source and commercial desirable

Jon: voice vote to accept results - unanimous

RESOLUTION: Members accept the results of the election

Operational Accomplishments

Growth - 58th member signed this morning

More members looking to join, esp on mobile, device manufacturers, business intelligence

going international - more europe, asia. 6 inquiries from Italy on first day

comment: accomodate feedback, changes to IP to allow others to join

jon: will discuss tomorrow

marketing committee

logos - web site - whitepaper

boloker: ran web site past analysts before press.

joe becker (IBM): press launch very well received. several leaders in press are eager to hear about OAA progress. ajaxworld coverage - on cover - most significant ajax news of the year

boloker: jeremy geelan (sys con) would like to share space with us in April NY show, more press

jon: a few companies got professional writing staff to improve white paper. pdf 22pp

jon: eclipse atf session yesterday. question on oaa white paper single/dual dom model.

becker: 10-20 press interviews with web launch. ~25 articles/reprints of pr.

chone: oaa must be inclusive

q: is there a member logo for web sites?

a: not yet approved. topic for tomorrow

openajax technical overview - coach

next steps

jon: expects initial checkins within days. iterations of adoption in major toolkit vendors. 1% work dropping code, 99% integrating

jon: target - widespread deployment early 2007

q: is license decided?

jon: ip issues covered in members agreement. Apache v2. original ownership, etc.

q: adoption must happen quickly. what are incentives? interop example?

jon: haven't figured it out yet, but some conformance test bed seems like a good idea. Need to discuss.

boloker: other early technologies. soap had an interop festival with 15 companies. validate hub, grows references.

jon: next meeting could be an interop gathering, in part

boloker: prefers sooner rather than later. small group sooner, everyone at next meeting?

chone: define compliance, provide an icon?

jon: need technology plus marketing hammer. good goal. how?

russell: test suite and compliance key to legitimacy long term. self-service

jon: test suite? sample app? pet store?

hakman: strategic recruiting would be helpful, e.g. test tool providers?

jon: terms- conformance? compliance? compliance hits "hot buttons". conformance seems better.

boloker: 3 todos. interop test harness (coach). marketing (jon). steering comm to use logo for conformance and what happens if misused.

q: is there a release policy?

jon: technical committee makes recommendations for policy, steering committee approval, adopted by open source project. Jon and James Margaris manage. Alex Russell and Lindsey Simon are committers. Other committers will be admitted based on merit.

jon: propose we discuss what we hear from customers

murray: customer to present running late (non-confidential)

break

vendor presentations on customer requirements

  • 11:00-12:30 - Ajax Customers Panel
    • Moderator: Erwan Paccard (ILOG)
    • Presenters/panelists (likely to change all the way up to meeting time):
      • Dave Stanek (American Greetings)
      • (Invited Guest) Jackson Thompson, Director of IT for Topaz Hotel Services
      • (Maybe we will have to have vendors report on what they are hearing from customers)
    • Panel participants are developers who are attempting to use Ajax technologies (i.e., not Ajax technology providers). What challenges they face and what they need from OpenAjax Alliance. It is likely that some of the people on the panel will be (non-member) invited guests.
    • Short presentations, followed by general Q&A

kevin hakman/tibco

acquired general interface 2 years ago. just open sourced. customers like hr block use multiple toolkits (Zimbra w/TIBCO) Skillset optimization is a problem. Work with and extend in a uniform way rather than fragments. Mashup toolkits like data.

much confusion around ajax term (lehman bros.) aggressive, lots of education. advises that ajax is not a single thing. it's a toolbox. use dojo for some things, dwr for others, etc. ecosystem, not one tool does it all.

versioning problems with browsers, readers. providers need to be part of the solution (e.g. jaws) hacks cannot be the solution.

accessibility: must have for some customers. In US, it's a competitive advantage. 508c is best effort (no tests for compliance?) In UK, there is liability.

john crupi/jackbe

based in dc area. intelligence customers. deployments are unknown. some jump from airplanes with bad connectivity. jackbe started in mexico where there is low bandwidth.

customer demands multibrowser/multi sources of data feeds. open stds. tell them ajax isn't really a standard but a set of standards. helpful to show oaa support from different size vendors working together. don't want to be locked into any one platform. security requests include RSA encryption on the client with DES, Blowfish. Had to write our own. Customers don't want the SSL approach for scalability reasons. Intelligence community says SSL is not secure enough, wants to deploy other encryption schemes.

Russell: TLS can choose different schemes

how do we do federated identity and authentication, connecting into web services. Customers want WS-Security but don't know why.

Russell: requests for smart card integration?

Crupi: no.

No .NET integration requests. Some in audience do.

Chone: because .NET users aren't using AJAX?

One customer said Atlas not as mature as other AJAX. Unsure where Atlas/XAML are going. Rumor is that MS does not know either.

Enterprise customers want to hear strategy. Use heavy middleware and Java.

No mobile requests. Good opportunity, but not seeing a play.

Portal free portals. Intelligence community using them to rapidly collect data, as a briefing tool. Traditional method was a powerpoint presentation, which provided no sources.

Jon: Interop requests for toolkits?

Crupi: haven't come across such requests, but customers don't want lock-in.

Anil/Vertex

product is wiki for business process documentation. Customer doing online printing. Moved from .NET to Java server-side on request. Challenges - customers don't want to change their HTML pages. Integrate existing HTML pages, scrape data, move to XML data. AJAX programming model quite a bit different from conventional programming. Lots of education. Integration- created a data model through to UI. Data source tied to HTML-based user interface, used by AJAX application.

Coach Wei/nexaweb

product deployed to 5000 enterprises. #1 issue not technology. most common question "are you ajax compliant?" "Do you support this particular toolkit?"

Hear a lot about offline computing, with desktop integration. AFLAC: 25,000 agents. Nexaweb provides desktop client with synchronization of data.

Many requests for .NET client support. Mostly within US. Europe extremely friendly to Java. Financial service companies have standardized on .NET for desktop.

accessibility: not much, except in public sector

Skill sets. SI in Japan community too small.

Scalability/reliability, particularly in context of real-time push apps like trading systems. Has to use HTTP, standard connections. Has limitations. Apache caches messages so pushes don't work. 2 CPU machine, standard Apache configuration can handle ~2000 connections. Need to deploy 100k or more connections.

Jon: how many have business requirement for offline? (~10 hands go up)

Russell: Dojo has offline demo. Uses some flash magic.

Wei: Need first page offline. Can't get it through the web.

Hakman: Uses IE cache w/offline button.

Wei: more offline requests on laptop than on mobile devices

Wei: Java Web start not liked. JVM dependency. Programmingmodel is challenging. Too big. Customers want instant on.

Q: i18n requirements?

haven't been asked about right to left. input methods, encodings.

hakman: has examples in arabic

topaz hotel services

central reservation service for independent hotels, runs web sites. button for reservation on smaller site takes you to another site. user might follow a long chain only to find out room is not available. made ajax widgets which can be embedded into customer sites (e.g. calendar with updates) to provide feedback, integrate booking process instead of having two separate entities

dev tools: notepad, vi. Started last September, rolled out in June. Looked at Dojo then, wasn't ready. Will probably refactor to use Dojo in the future. Only off-shelf component was prototype. On back-end, PHP (REST) web services. Best tool is Firefox JS debugger. Zend,Macromedia,Eclipse do not do what you'd expect from an IDE. Dealt with browser incompatibilities. If AJAX support not available, fall back to old style page, no AJAX components.

Google Map mashup with availability

20th anniversary. First on the Internet to do direct reservations in 1994.

Flash engine is popular. Done with movie making model. Looked at Laszlo. Chose AJAX over those.

Support IE5+ FF/Moz 1.0-1.5, Opera

DB: end of life informix

Some lob is green-screen still

Q: did you try a degradable interface? (no js)

no. use separate pages.

Q: how many users can run AJAX but don't know it?

conversion rates going up. users don't know or care about AJAX, or if it's Flash. As long as it works.

technology, open ajax is good message to corporations. but keep transparent to end users.

50% find the property but end up calling on the phone. credit card security?

enterprise java/jsps? speaker worked at Sun/Netscape (iplanet) some java, but php won, used java where php wasn't able do something

new interface is ~2% increase in conversion rate. good, but not huge. summer conversion rates are higher anyway. don't really have enough data.

scalability concerns: small shop, not much of an issue. would just go buy another $1k machine. servicing ~300 hotels on sf consumer site. No chains with >10 properties

Q: decrease in call center inquiries?

don't know.

focus on niche market, excel at customer service.

expedia, hotels.com have models which aren't as beneficial to the hotels. reduced margin.

american greetings

... showed up late. sorry

only works with js turned on

Flash 6, which sometimes requires users to upgrade

92% explorer, 6% firefox, 2% other

runtime panel and server-side panel

  • 1:30-3:30 - Ajax Run-time Panel
    • Moderator: Jon Ferraiolo (IBM)
    • Presenters/panelists (likely to change all the way up to meeting time):
      • Someone from GI team (TIBCO)
      • Someone from Spry team (Adobe - at risk - Adobe may not be able to attend)
      • Maybe Chris Matthews (ZOHO) on component architectures
      • Someone from XAP project (Nexaweb)
      • Anil Sharma (Vertex Logic)
      • Maybe Andre Charland (Nitobi) on component developer issues
    • Panel participants are Ajax technology providers that are delivering pre-packaged client-side JavaScript libraries, JavaScript components or web browsers. What challenges they face and what they need from OpenAjax Alliance.
    • Short presentations, followed by general Q&A
  • 4:00-6:00 - Ajax Developer And Server-Side Tools Panel
    • Moderator: Christophe Jolif (ILOG)
    • Presenters/panelists (likely to change all the way up to meeting time):
      • Christophe can talk about ILOG's use of Ajax
      • Chris Shalk will present on Oracle's vision about Ajax and OpenAjax
      • Jochen Krausse will present Innoopract's vision about Ajax and OpenAjax
    • Panel participants represent Ajax technology providers who deliver complete products for developers, where the product isn't just a run-time library. Examples include server components, server-side transcoders/compilers and IDEs. What challenges they face and what they need from OpenAjax Alliance.
    • Short presentations, followed by general Q&A

Chris Matthews

JWAX.org dynamic js building. common component architecture, inheritance relationships. html render kit. working on a dojo render kit. might be arranged per browser or per component.

Come up with an architecture for openajax for components which are aggregatable, OO inheritance, etc.

kevin hakman

javaone. demonstrated how to use toolkits together. pleased with alex's code.

takes too long to get things through w3c. apis and declarative markups with openness to new components, managed.

unit/load/stress testing concerns from customers. successful with selenium. talked with mercury. strategic recruiting. we have toolkit vendors and users. need whole ecosystem

anil sharma/vertex logic

Slide deck: Image:VertexLogic OpenAjax 20061005.pdf

wish list (Data Source/ Model Specification)

Universal client can connect to any data source and transform it to info provided that the data includes enough meta data to tell UI "what it is". A step prior to that is - client converts data to some generic info, user augments it with additional meta data by changing UI to edit mode.The above two statements are futuristic. But the point is, UI is all about understanding meta data of application data and creating user friendly views of it.

We can come up with a higher level specification of model that can generically be used by any UI toolkit. It is possible to define a minimal meta data (without making it overly complex with MDA, complex Schemas, domain models etc). Such a set will include - for each object certain well know attributes such as UID, class name, version, position of object in a collection. Minimum meta data will include class name, applicaiton specific key attribute, associations. The above info is enough to represent an Object, Object Collection and Object Tree (of associations) in a generic way. It will be a big step in standardizing data source. This will work well with the dynamic programming model supported by AJAX.

Kevin: Is it similar to SDO?

A: Yes, in many ways.

alex/dojo

dojo background, history

dojo is layered library that javascript never had. the 'dojo way': reduce barriers to adoption. simple to use first, fast later. liberal licensing.

bias towards html, but dojo can run in any js environment.

q: why didn't JS have this library built-in

prototype functional language that looks kind of like C/Java. because it doesn't have a class library, it's very light, simple, and ubiquitous.

would be good if we could get out of the business of certain things (e.g. widget system) and get browsers to do this. meanwhile, we provide shims. hope we can drop them later.

XAML - serialization of .NET runtime. To host XAML, you need to implement .NET framework, full toolchain. Non-flex, non-MS are betting on open web.

q: how does Dojo build widgets dojo's philosophy on widget: automates markup conversion to DOM, hidden by templating system with parser. turboajax is a company which builds dojo widgets - example of commercial, open source partnership.

dojo tries to be a level playing field. anyone can take, contribute.

q: browser stats on dojo web page see /cgibin/awstats.pl 46% firefox/53% IE. Had been as high as 60% FF. IE 7 is extremely low (2%)

greg/sun

jMaki (maki means wrap) Started 1.5 yrs ago. http://jmaki.com/about

wraps the different toolkits out there (dojo, google, spry, scriptaculous, yahoo...) programatically creates a widget, feeds data. Can come from JSP tag / JSF component. Scriptlets with Javascript can be used to inject widgets.

to load dojo, turn off widget parsing, since jmaki is already aware of the tags supports postconditions for loading complex widgets.

basic pub/sub mechanism (handled in event hub) e.g. map and geocoder widget, bind properties together without using any specific vendor's event bus. very similar to dojo, but on the jmaki object.

other challenges: jsf components, pet store. needed javascript and runtime packaged, avoid multiple copies.

different versions of libraries is a problem also

a lot of java developers are afraid of javascript,css. a tag is not as scary

christophe jolif/ilog

Slide desk at: Image:ILOG OpenAjax 20061005.pdf

Members Agreement discussion

jon: adobe had reservations about signing member agreement, but wants to work with alliance to figure out how to address concerns. jeremy chone will discuss:

chone: fair for us to give patents so that a company doesn't hold working group hostage to a patent. but don't want to force companies to give everything they have if a new working group is formed.

jon: yahoo may have issues internally which take time. not clear if it's a problem with the member agreement.

dietzen: zimbra raised same concern. cynic: just make a working group and get another member's IP. recommend an opt-out of openajax, no implied retroactive grant of ip for anything that might ever come up in openajax in the future

jon: options - change members agreement on a per-working group basis (adds layer of management) another approach - notification, may not require members agreement change. steering committee could adopt policy such that any time ip is broadened, such as forming a new working group or spec, there would be a notification. if oaa takes on a lot of activities, it will be impossible to monitor.

chone: adobe would prefer opting in certain committees and choosing where to donate ip

q: being 4/5 part of the open ajax alliance might not equal compliance?

wei: how does jcp handle this issue?

david: transfer responsibility to sun (per jsr/working group)

wei/dietzen: smaller number of working groups means big companies reap ip whether or not they are a member

chone: oaa will benefit by having big players (yahoo, google...)

dietzen: likes the opt-out, quit open ajax option. not having that option is unfair

wei: in apache, if a company goes against license, they cannot use apache product. in openajax, organization has to leave alliance and cannot use anything produced by the alliance and would be infringing on the patent.

dietzen: doesn't think that's possible

chone: basic problem is that adobe does not know what they're giving up joining the alliance

coach/nexaweb

Introduction to Apache XAP (charts)

declarative model to build and simplify apps


andre charland

high performance ajax components. scrolling data grid, keyboard navigation, copy/paste with excel docs

would like to see vb-style development, best practices for working with frameworks and ides

Operational issues

Friday Oct 6

  • 9am-10:30a - Open discussion about operational issues
    • Moderator: Jon Ferraiolo (IBM)
    • Topics should be proposed before the meeting begins. Here are some candidate topics:
      • How many all-hands meetings per year? 1? 2? Always North America, or take turns geographically?
      • When/where should next meeting be?
      • Should Alliance have a confidentiality policy regarding committee discussions, wiki, and mail lists?
      • Should we allow individuals to join (vs companies)? If so, how should this work?
      • Voting process - any changes for next year?
      • Switching to yahoogroups for mailing lists
      • Who do we need to recruit and how do we get them to join? (For example, more open source runtime libraries? More open source Java projects?)


Propose: Meeting like this one in April around AjaxWorld in NYC, IBM can probably host. Apr 4-5, 2007? One day membership, another day interop fest?

Milinkovich: how often? 2x/year? notice can be 60-90 days?

Jon: Justification for one day: more about information exchange than decision making. Resolution? Ok.

Jon: try to align with some show other than ajaxworld next october?

Jon: minutes are confidential

Russell: why?

Milinkovich: follow open source rules of engagement, be transparent. starting now.

Jon: Nothing prevents a member from forwarding private wiki data to the public

Milinkovich: why is information not just public

Hakman: is it an infrastructure (wiki) acl issue? probably not

Milinkovich: anyone think there is a need for confidentiality?

(no response)

Wei: Is anything at Eclipse confidential?

Milinkovich: working group activities are all public and transparent.

Russell: review process?

Milinkovich: Eclipse has never signed NDA Some board meetings not public Abridged steering committee minutes might be ok

Wei: W3C many conversations are confidential, limited to members

Milinkovich: does not want to follow that

Simon: Ok to publish IP issues from yesterday, for example (Adobe) We don't want everyone to blog about such a thing

Milinkovich: reasonable to say that concerns were raised, but keep wording neutral and not give details

Jon: people can go off the record as the exception

Russell: Dojo would not be okay with meetings that cannot be shared. Conflicts with Dojo policy.

Jon: consensus on openness

Dietzen: assuming everyone is reasonable and uses discretion

Milinkovich: ground rules should go in meeting agenda

Jon: Action: make wiki available to the public, read only

Jon: resolved, mailing lists should only be postable by members, archives are public

RESOLUTION: Wiki and mail lists are read-only to public

Milinkovich: is there a mechanism for public to address an issue?

A: could find a member of the working group (email addresses are visible)

Milinkovich: or have a separate list (oaa-public? openajax@yahoogroups?) requires pull rather than push (someone has to monitor list)

Hakman: try it and we can always change mechanism later

Jon: resolved, a single public openajax forum.

RESOLUTION: Single public openajax forum

Jon: proposal, ask our isp to get mailman working, otherwise switch isps. yahoogroups has a delay, not good for discussions. unanimous agreement.

RESOLUTION: Ask our isp to get mailman working, otherwise switch isps

Q: use ajax with mailman?

Dietzen: might just alienate others in community who compete

Milinkovich: Eclipse remain vendor neutral all the time

Jon: just use normal interface for now

Q: use ajax on openajax website? same issues.

Russell: mailman is a FSF project

Jon: Voting process comments?

Hakman: submit votes as a numbered list rather than ordered list to enforce ordering

Hakman: public description of scoring algorithm useful (in the wiki somewhere?)

Jon: We use same algorithm as Eclipse. Not really that complicated

Q: might voting be used for issues as well as elections?

Jon: always possible, e.g. change member agreement. Otherwise, consensus is used most of the time

Milinkovich: +1/-1 mailing list voting for issues is common, works fine.

Jon: concerned about conspiracies - individuals voting for steering committee, where a corporation could register individuals to skew a vote

Milinkovich: each company has one vote

Russell: If it gets that bad, we might as well start over

Milinkovich: is it in the interest to extend membership to individuals

Jon: might have different non voting status for steering committee, like w3c does

Russell: some open source projects are one person.

Jon: has 20-30 requests for indiv memberships

Milinkovich: how many would be ok with simply public mailing lists?

Jon: maybe half; but some want to contribute. Some might be at a university and can't get approval for a small project

Milinkovich: meritocracy. Only open to people who bring value. perhaps by invitation

Russell: openajax is not an opensource project, so this may not all apply

Milinkovich: by contributions to public mailing list, etc., members of working group might extend invitation/vote?

Jon: committee proposes. steering committee approves.

Proposal: individuals can join. working groups must propose to steering committe which votes. nominations based on meritocracy.

RESOLUTION: Individuals can join, but based on contribution and meritcracy. Working groups must propose to Steering Committee which must accept.

milinkovich: need terms of use on website so that ideas on mailing list are licensed to the group, can use Eclipse as a template. on every page.

Jon: action will take care of terms of use issue before all website info is made public

'ACTION: Just to ensure that Terms of Use are available before making more information public.

Jon: recruiting. need to contact scriptaculous/prototype. others? DWR

Hakman: test suite vendors. Mercury, selenium, testng

apache? kde?

mobile.

Jon: heard from symbian, hyperion. big and small.

Russell: whatwg?

Jon: already through w3c

hixie/hyatt/hakon/brendan...

Jon: most orgs are already members

Russell: whatwg might be a good backdoor to other orgs

Jon: another topic - what we can do to push browser vendors

-break-

Marketing Committee Breakout

  • 11:00-12:30 - Breakouts - Marketing Committee and Technical Committtees
    • Introductions so everyone can associate faces with names
    • People make short statements about recommendations for future work within the committee
    • Conduct a regular committee meeting, addressing current topics
    • Marketing Committee topics:
      • Driving the industry towards interoperability: OpenAjax "compliance" vs "conformance" vs ???
      • Roadmap and priorities, particularly white paper(s)
      • What to do about "compiled ajax" category
      • Adding TIBCO/GI to list of open source projects in White Paper
      • How to integrate the FAQ?
      • Logo usage rules like W3C? http://www.w3.org/Consortium/Legal/logo-usage-20000308.html
        • Some members want to put OpenAjax logo on their Web site. Is this OK? Which logo?

Attendees

  • ILOG : Erwan Paccard
  • IBM David Boloker
  • OpenWave: Dan Nguyen
  • Eclipse: Mike Milankovich
  • Software AG: Shane Mcmillan
  • Natobe: Andre Charman
  • SCALIX: Florian von kurnatowski
  • ICEsoft: Chris Ericksson
  • SUN: Jim Driscle

Web site:

  • All: We rename Comities to Working groups
  • All: Clean the member list in 2 weeks to let large corporation go through the approval process. We no more list participants but only members.

- Discussion on possibility to opt-out based on yesterday’s Adobe feed-back on IPR

  • will be lunch discussion for the steering committee
    • we need to list who we believe we need to be part of the alliance and adjust the IPR if needed
    • overall people agree that even if strict existing IPR is a strong commitment fine to us
    • Erwan: let’s be pragmatic, if IPR prevent key players from joining, let’s be more flexible
    • David: then anyone can opt in & out and you end up with a hardly manageable organization
    • David asked for opinions.
  • All: no reason to change now, let’s discuss it again if Google/Yahoo/BEA are not willing to join because of that

Ajax Technical Whitepaper

  • David: do we extends it? Mobile / security /

Ajax MKT whitepaper

  • Chris: we need to clarify value & market segments.
  • Erwan: we already have a draft, let’s revive it
  • All: we need such a paper on "Why AJAX". Chris agreed to help push the MKT white paper to the next step.
  • David: we should also take care of AJAX pitfalls / issues and approach to adoption
  • Mike: we need to address 2 hurdles: how to get enough developers fluent in AJAX and how ???
  • Florian: we need SWOTs to position AJAX against alternatives
  • Chris: sensitive to position against other RIA approaches even though it would make paper more complete
  • Dan: we have to because RIA users will want to have answers
  • David: need to get lawyers opinion, dangerous path, can easily be suied
  • Erwan: I would not this path, positioning AJAX against other RIAs approaches
  • All: we do not position against other RIAs, we just mention for completeness

OpenAjax interoperable "Mark"

  • David: how do we handle it?
  • Mike: technical committee will develop and open source specifications and test harness
  • Mike: we need fast a trademark policy public on the web site
  • All: agreed
  • Chris: Member logo is easy and interoperable logo is needed
  • all: we cannot validate, we do not have resources
  • Chris: let’s simply to self-certification with freely available test harness
  • Erwan: like the self-certification and simply having a page listing the ones we know and tried suickly could become a market reference people refer to
  • All: we cannot afford more than self-certification process
  • David: the market will fix itself if someone claims to be interoperable but is not
  • Mike: later down the road, we will have more customers part of the organization than providers. Customers are the ones looking for certification and they will take care of adjustments if needed.

Recruitment

  • David we need to proactively open to consumers companies, other open source projects and people in other geos
  • Mike: the meritocracy process for individuals will be important too
  • David: many of those large corporations do not do AJAX because of accessibility
  • David: then adding security / localization / … to the list of their concerns

Discussions to continue next Tuesday

  • Open source projects
    • FAQs

Technical Committees Breakout

    • Technical Committee topics:
      • Roadmap for next few months
      • Keep Interop and Markup as separate committees or join together?
      • What happens in committee meetings versus what happens in open source project?
      • Assignments and commitment dates for Hub activities, such as integration with toolkits and products
      • Where are Hub docs developed? openajax.org public web site? sourceforge public web site? member wiki? within sourceforge project trunk/docs?


People introduction

CW - Coach Wei (Nexaweb)
            - client side ajax - xap
            - ss java
            - involved with eclipse atf
            - Areas: lwo hanging fruits
                        - evet hub
David Sinick (American Greetings)
            - web apps several years, python
            - java, php, .net
            - Would like to see: interop, potentially new libraries, some work there 
 
Andrei Dragomir (Adobe)
  - web apps about 5 years, client side mostly
  - architect / main developer for ajax projects
  - Would like to see: interop, porentially best practices proposals
 
 
Christophe Jolif (Ilog)
            - advanced visual components - java, .net
            - on top of the jsf framework
            - Would like to see: how ajax integrates with ss components
            - Would like to see: interop, we have to integrate with others
 
Rick? Evans (sun)
            - finished a course on creating ajax enabled jsf components
            - Would like to see: promote best practices, interop
 
Anil Sharma (Vertex Logic)
            - business process documenting tool
            - Ajax about one year
            - ui dev swing, .net, struts, java ss
            - Would like to see / Could do: contribute in coding
 
Lindsay Simon (Finetooth)
            - framework development script.aculo.us
            - Would like to see / Could do:help with interop, guidelines for people to think how to do that
 
Joonas Lehtinen (IT Mill)
            - dhtml 8 years
            - ss java development toolkit
            - Would like to see: coding conventions, best practices
 
Jon Ferraiolo (IBM)
            - IBM from may (adobe before)
            - solely dedicated to help with OpenAJAX Alliance
            - church vs state
            - community stuff
            - Menlo Park ( down the road)
 
Adam Peller (IBM)
            - emerging tech at ibm
            - QEDwiki project
            - contributing to mozilla
            - contributor to eclipse atf project, dojo toolkit
            - Would like to see: where standards might apply, best practices for tooling in ajax
            - Would like to see: accessibility initiative
 
 
Alex Russell (Dojo Foundation)
            - dojo toolkit
            - sitepen ( consulting company)
            - interest: push the web forward, into any device existing
            - what we can do: put some teeth behind the interop stuff
            - Would like to see: dialog with browser vendors is useful, make them understand our needs
 
Steve Mariyka (...lost the name)
            - approach JSF
            - minimize the amount of javascript required in these solutions
            - they always run into situations where people want different stuff
            - Would like to see: creating a component server architecture, to integrate in the jsf model
  
Chris Schalk (Oracle)
            - product management
            - experience with db, midle tier tech
            - currently JSF, java technologies
            - oracle would get to technolgy to work, not get left behind
            - Would like to see: formal declarations approach how to create ajax widgets / components
            - Would like to see: how other people view the component apps approach
 
Scott Dietzen (Zimbra)
            - 135K lines of javascript
            - no html, challenges
            - zimbra used as a benchmark for browser
            - use a stuff like qtp
            - open source benchmark project 
            - convince the browser vendors to test that, people can see the result of running ajax
            - Would like to see: layered functional test, establishing different levels of ajax support, influence companies to become better at supporting ajax  
            - hard to deliver ease of use and integartion
                        - better job at integrating, picking some tools that work good together
 
Chris Matthews (E Link Biz)
            - jwax.org
            - salesforce automation company
            - erp, supply chain management
            - componentization, aggregation, reuse
            - Would like to see: component reuse, defferent matching of diffferent tools and lbraries
 
Greg Murray (Sun)
            - creator of jmaki
            - different solutions of doing java and ajax
            - Would like to see: basic functionality to be put in the browser
    - component model: too far for now
    - some conventions for start
    - OO orientation
    - have already did some os conventions for sun
    - default java objects
    - update the json components
    - security (in the future)
 
Kevin Hakman (tibco)
            - cofounder GI
            - deliver ajax interactive rich clients
            - toolkit swing (2001)
            - visual toolkit overlay
            - build ajax in the browser
            - committed to the open source movement
            - Would like to see: interop for dojo, yahoo, google components
            - guidance, is needed, to direct the chaos ( positive)
            - not get bogged down in an extensive formalization priocess
            - Would like to see: markup scanner - very important near term objective
 
Sam Silver (OpenWave)
            - web messaging apps group
            - more compatibility across diff. toolkits
            - Would like to see: get the browser manufacturers to establish some standards

Agenda (jf)

            - the primary focus: complete the open ajax hub
            - standardize component wrapping
            - jon: big fan in getting into serverside
    - if we have time for other things, we would do it
            - CS : just mentioning the components to put things into context
            - SD : couple of concerns; we don't have enough of a roadmap for non-expert developers to go ahead and have fun with ajax
                        - doesn't bring ajax to the power builder crowd
            - KH : strategy for enabling components to generate from simple markup
                        - some activity in the sense of promoting best practices
            - AR : not a new tech, a new id way of doing things. we lack someone from the interaction design perspective
              SD : hitting the oo architects, programming skills. not putting it together
              CE: are we hiding javascript from the java / web designer developer
              KH: not java mentality, but programatic vs a tag based approach
              CW : the 4 established components
                        - loading / unloading
      - collision
      - event management
      - markup scanner
                        CS : jsf component first, then ajaxify, then hide the js from the end user
                        JF : solve the interop stuff
                                     can't just do client-side
                                     should always do also server 
                        SD : lose too much stuff, lose from sitght the gaps. 
                        SD : shouldn't just focus on one thing
                        JF : integrate with ides, anything more than that
                        - bring the problem / solution up the stack, tooling
                        - describe what we have been done
                        - promote ??? Eclipse ATF
                        - json description of components
                                    - atf consuming, tibco consuming the descriptions
                        - jsf integration ? 
                                    - driven by jcp, we may do only 20% of the work. 
                                    - contract with the client side
                                    - define a format and a protocol
                                    - should we do it ? on the first batch of the output ?
                                    - tons of work for the event manager / hub
        - different def of success
          - backplane for events
          - just throwing them around
                                    - test
                                    - getting toolkits to adopt the test/hub
                                    - markup scanning: giving it up to whatever toolkit
                                    - oaa reserved namespace for event hub
                                    - should see who wins
      
            plans: work on hub and test
                        - roll it out in to toolkits, just as soon as it is stable enough
      - toolkits and selected ides

IDE integration

    - make sure that we ar doing things with ides properly
                        - eclipse atf, doesn't work with generic markup right now
                        - json / xml potentially
      - just a standard
      - json, subset of xml
                        - the hub is orthogonal to the description of the markup
                        - in order to be useful, must be consumed by people
                        - ACT: greg: wiki page for a proposal about the markup / metadata description
                        - ACT: get the ide teams determine what they are working on
                        - ACT: server guys check to see if that is integrated with their approach
                        - component that also have default data
                        - makes sense to make the two groups into the interop group
                                    - break that into some subcommittee
                                    - browser bashing/ toolkit bashing
                        - committee = multiple task forces
                        - single technical wg, taskforces
                                    - that work on projects
                        - initial kind of task forces ?
                                    - jsf
          - rick, greg
          - chair: Chris
        - ide integration
          - mike, kevin
          - chair: mike
  
                        - SD : we sould also address the best practices / benchmark stuff
                        - benchmark: nobody has any comprehensive benchmark
                        - maybe a set of pages to test
                        - ie: nodes iteration
                        - SD: should really have a reference app, the equivalent of hello world in ajax
                        - wednesday 8.00 (pacific time) for the phone calls

RESOLUTION: Merge Markup committee into Interoperability committee

RESOLUTION: We will have a Server-side/JSF task force manned by Rick, Greg and Chris, with Chris Schalk as chair.

RESOLUTION: We will have an IDE integration task force manned by Mike and Kevin, with Mike as chair.


Mobile Ajax

  • 1:30-3:30 - Mobile Ajax - what should we do about it
    • Moderator: Jon Ferraiolo (IBM)
    • Presenters/panelists:
      • Jon Ferraiolo (IBM) - Introduction about interest expressed by several companies and importance of mobile devices going forward and comments collected from companies unable to attend this meeting
      • Alex Russell (Dojo)
      • Someone to talk about JSR290 (Sun)
      • Christian Sejersen (Openwave)
      • Håkon Wium Lie (Opera)
    • Short presentation of position papers by selected members, followed by general discussion

Demos from Jon, Alex Russell, Christian Sejersen/Openwave, Hakon Wium Lie/Opera&WHATWG, Nandini Ramani/Sun

Hakon demos: Google suggest (asynch), Canvas w/audio, Javascript + 3D engine

Nandini demos: tools (NetBeans IDE)


-break-

Marketing/Technical activities for Open Ajax (Jon)

Mobile working group? Vision + interop (browser bashing/test suites)

Jon: subsetting? target open mobile alliance (OMA) stack

q: AJAX hub should work on mobile devices

a: yes, but it's simple javascript (no dom, no eval...)

q: some vendors aren't focused on javascript

jon: yes, javascript is the purpose of the alliance

sejersen: oma has tests. oma does not have xhr support. oaa might point to list of specs, write tests for that. easier to promote if we have a spec, but can be done by referrring to other specs.

lie: one www. specs should be same across mobile/desktop. shouldn't be a mobile profile. don't wait for market to decide.

sejersen: same app will not work the same way on device with low performance, slow network, little memory

?: not unique to mobile; desktops can perform differently

russell: browsers always won because of loose contract. content can be rendered in different ways, degradability. web developers know this & don't expect fidelity. this is not a problem.

sejersen: despite openajax certification, apps would not perform reliably across devices

milinkovich: guarantee failing gracefully?

hakman: tibco might use different set of ui controls

russell: ajax covers both applications and upgraded pages

jon: mobile ajax under standard definition hasn't happened today. 3% of devices isn't enough

milinkovich: will happen over time. what can we do to accelerate? killer app?

jon: focus on native issues? e.g. no mouse, limited input device

hakman: test suite must support mobile

milinkovich: not hearing anyone demand the activity or offer to lead it

  • Decision to not form a Mobile Working Group at this time because we do not have a proposed charter at this time nor a list of people who would participate. Instead, at least until such a proposal is received, we will ensure that existing activities are mobile-aware and include mobile issues within their activities.
  • General leaning towards adopting a "one Web" approach where we do not promote subsetting for mobile devices. Rationale is that Mobile Ajax seems unlikely to take root on a large scale around mobile subsets and since no one is volunteering yet to lead Mobile Ajax interoperability around a subsetted Ajax.

Steering Commmittee

Jon announced results from Steering Commmittee meeting over lunch.

RESOLUTION: IBM/David Boloker will chair the Steering Committee

RESOLUTION: Meetings will occur every two weeks, agenda will be posted one week in advance along with any documents to review. Voting on issues can be done electronically.

RESOLUTION: Mike Milinkovich will pull together a proposal for process for creation of Working Groups and advancement of Specifications such that Members are notified and have opportunity to opt-out of the Alliance

ACTION: See previous resolution.

RESOLUTION: Before formation of a Working Group, the Working Group must have an approved charter that defines the scope of activity.

ACTION: Jon to write up initial charters for Marketing and Interoperability working groups.

Futures / Wrap up

  • 4:00-6:00 Futures discussion and wrapup
    • Moderator: Jon Ferraiolo (IBM)
    • Presenters/panelists:
      • Alex Russell (Dojo)
    • Short presentation of position papers by selected members, followed by general discussion
      • OpenAjax conformance - how to define, how to motivate industry, do we need test suites or enforcement mechanisms?
      • Ajax security?
      • Requirements for portals, wikis, and other mashup scenarios?
      • IDE issues?
      • Client-server communications - should Hub coordinate communication requests?
      • Standardize popular APIs (e.g., mapping)?
      • Drag&drop?
      • Data binding?
    • Wrapup - Jon Ferraiolo

conformance vs. compliance: leave as topic for marketing committee

security

ajax security: infoworld article this week concerning flaws. best practices?

russell: better sandboxing, key management, crypto

boloker: authorization & access control

hakman: fud vs. reality?

russell: mostly fud

boloker: old, well-known problems. enumerate problems, solutions in a working group.

addresses fud.

russell: cites crockford's work on json cross-domain parsing

volunteers: dojo, sun, ibm, tibco...

russell: owasp, mitre, cde

RESOLUTION: Form a Security Working Group to enumerate problems, propose solutions, and develop documents for the community

technical task forces

jon: requirements for wikis, portals... addressed in interop. separate task force for ide. hold off on mashups.

russell: portals and mashups imply versioning problem

message hub

greg wilkins: jde web server, communications. working on dojo/cometd, jms on the browser, working with dwr. fundamental problem: ajax is http request/response, not a good foundation because, in comet, connections are a limited resource. (2 active connections per browser) Ajax push: 2 frameworks exhaust resources. Efficiency issues- event triggers multiple requests with round-trips. Makes page very slow. How do we control access to this scarce resource? address messages so they don't collide? jms-like communications protocol in dojo/cometd. might need something less to send messages and target at a component or framework.

wilkins: more resources would only make problem worse, stressing servers

proposal for message hub?

RESOLUTION: Greg will author a proposal for message hub, possibly leading to a task force or new feature for OpenAjax Hub

data binding

data binding: may require hub or binding facility, but not right now

russell: dependency on ie implementing getters/setters

hakman: data layer is a convenience, value add for adoption, but not a hard requirement