MINUTES OF PAVR INTERIM PROGRESS MEETING
Battelle Research Centre, Geneva, June 4th, 1997

SUMMARY OF ACTIONS ARISING
PJW
  • Establish UIB's involvement in project
  • Obtain EC's formal approval to revised workplan [DONE]
  • Chase Siemens on their role in achieving D17 by end of year 1 and inform J-DF
  • Contact Siemens re possible visit with or by J-DF in July
  • Check TMR rules on applicants from associated countries (in this case Israel) and inform J-DF
  • Send J-DF Michael Brooke's and Alastair Kilgour's email addresses [DONE]
JB
  • Describe UIB's involvement in project
JP
  • Chase own contacts on J-DF's behalf for speech recognition candidates
J-DF
  • Check terms for multi-site licences for SoftImage
NMT
  • Check terms for multi-site licences for Wavefront and 3D Studio
WP Coordinators
  • Each to organize first technical meeting
All
  • Refer any suitable applicants (speech recognition) to J-DF (Ecole des Mines, Nantes)
  • Refer any suitable applicants (dynamic simulation) to J-DG (UJF Grenoble)

Present
University of Bath Phil Willis (chair)
University of Glasgow John Patterson
Ecole des Mines, Nantes Gerard Hegron, Jean-Daniel Fekete
EPFL, Lausanne Daniel Thalmann
University of Geneva Nadia Magnenat Thalmann, Pascal Volino, Laurent Moccozet
TU, Vienna Michael Gervautz, Francois Faure
UJF, Grenoble Jean-Dominique Gascuel
INRIA, Rennes Bruno Arnaldi, Stephane Donikian
UIB, Mallorca Eric Chauvinau
LUC, Diepenbeek Frank van Reeth

Apologies
Siemens, Munich

  1. Welcome, introductions and departure times
    Phil Willis welcomed those present, and new-comers in particular, and introduced Anna Beria, the project's part-time administrative assistant who would take the minutes of the meeting and be responsible for the project's routine administration. Constraints on departure times were duly noted.

  2. Minutes of previous meeting
    The minutes were adopted, subject to the report of Group 2 discussion being produced by John Patterson.

  3. Matters arising from minutes
    Point 3: Planned schedule for D0. UIB's involvement still not clear, as Eric Chauvineau, only recently appointed under the project, was unable to comment.
    § PJW to chase

    Point 4: Review of appointments: to be discussed under Item 4 of the Agenda

    Point 5: Project demonstrator: to be discussed under Item 5.

    Revised workplan details had been sent to EC. No reaction had been received, but no problems were expected. As formal approval was required, PJW would follow-up.
    § PJW

    J-D Fekete noted that Siemens' absence made D17 (Eye detection - due by end of year 1 under WP4) unlikely to be met.
    § PJW to chase

  4. Review of appointments
    TU Vienna was ready to offer a PhD position to Carol O'Sullivan from Ireland. However, she would be based in Dublin and it was not clear to which resources she would have access. M. Gervautz asked whether travelling expenses incurred by potential appointees could be charged to the project. PJW would check with the EC if such an arrangement was acceptable under the terms of the TMR programme, since mobility was the key issue. Francois Faure, from France, (who was attending the meeting) was another potential appointee.

    D. Thalmann referred to EPF Lausanne's complicated staffing arrangements, which included Tom Molet (PhD) working on motion capture and a second researcher (Amaury Aubel), but since EPF did not receive EC funding, this was not a problem for the group to address.

    F. van Reeth (LUC) had exchanged email with John C. Lombardo, from France, who was currently working with C. Puech at UJF, Grenoble. He would be free from November and was interested in a post-doct. position. Another French researcher had expressed interest and was being considered, though this seemed less likely to result in appointment.

    J. Patterson (Glasgow) was finalizing the appointment of Jean-Christoph Nebel (F) to a post-doct. post from beginning of November. Another Frenchman, Frederic Labrosse, had been interviewed, but a decision had not been made yet as he was still working in Montreal. JP pointed out that he was currently on sabbatical leave and this made administrative arrangements difficult and supervision virtually impossible. Things would improve next year.

    J-D Fekete (Ecole des Mines, Nantes) asked whether an applicant from Israel might be acceptable, as Israel had concluded an association agreement with the EU.
    § PJW to check rules

    No suitable candidates had been identified as yet and the search continued. The participants were invited to email EMN directly with details of any post-doct. or PhD applicants from HCI community interested in speech recognition.
    § ALL to introduce suitable applicants
    It was also proposed that J-DF visit Siemens with PJW in July or alone, depending on their being interested.
    § PJW to liaise with Siemens

    § PJW to send Michael Brooke's and Alastair Kilgour's email addresses

    § JP to chase own contacts

    The University of Geneva had already appointed two researchers: Pascal Volino and Laurent Moccozet, both attending the meeting.

    UJF Grenoble was still looking for suitable candidates interested in dynamic simulation. The participants were again invited to refer any hopefuls directly to Jean-Dominique Gascuel.
    § ALL to introduce suitable applicants

    INRIA had two available positions, 1 PhD and 1 post.doc., but no suitable applicants had been identified yet. Job specification was being drawn up. Frederic Beauchamp and Remi Cozot were already working on their chosen task (Collision detection).

    Eric Chauvineau, from France, was UIB Mallorca's newly appointed PhD student.

    The University of Bath had appointed the part-time administrative assistant and had a short list of three or four applicants for the research positions. One had already been interviewed by PJW and was appointable and a second one was a distinct possibility (see above). Another two had yet to be interviewed. The first appointment seemed fairly imminent.

    It was agreed to postpone any discussion on deliverables until the next meeting since the appointment situation was still too uncertain. PJW reminded the meeting that the first report to the Commission was due by the Autumn meeting.

    SUMMARY OF RESEARCH APPOINTMENTS (Vacant posts in brackets)
    LOCATION POST 1POST 2
    University of Bath, UK (Post-doc)(PhD)
    University of Glasgow, UK Jean-Christoph Nebel(Post-doc) [FR](PhD)
    Siemens, Munich, DE (?)(?)
    EMN, Nantes, FR(PhD)-
    EPFL, Lausanne, CH Tom Molet (PhD) [F] Amaury Aubel (?) [F]
    University of Geneva, CH Pascal Volino (PhD) [F]Laurent Moccozet (PhD) [F]
    TU Vienna, AT (Post-doc)(PhD)
    UJF, Grenoble, FR (Post-doc)-
    INRIA/IRISA, Rennes, FR (Post-doc) (PhD)
    UIB, Mallorca, ES (Post-doc)Eric Chauvineau (PhD) [F]
    LUC, Limburg, BE (Post-doc)-

  5. Project demonstrator
    This was the project's key issue, and at the Limburg meeting the participants had spent some time divided into two sub-groups to discuss how best to integrate the various teams' work. Group 1 report had been circulated as annex to the minutes. Group 2 report had met with technical problems. Since a decision on practicalities was now required, PJW asked the participants whether there had been any developments since the meeting.

    J-D Fekete reported that no further discussion had actually taken place after the meeting, but thought had obviously been given to the issue and informal discussion had already been resumed. There seemed to be some agreement on the need to test "SoftImage" and move to a PC platform. VRML 2.0 was too basic, MR Toolkit too expensive and Maya not yet available.

    J-DF had attended a SoftImage demonstration which had proved pretty promising. The package was affordable, at 5,000FF for 1 year and 10,000FF purchase price, could be developed to support both frame-based systems and volume-based applications, and had real-time capability. It was agreed that a copy would be obtained and evaluated by October. Alternatives should however be considered alongside SoftImage so as not to waste time should the frontrunner prove unsuitable after testing.

    Nadia Thalmann stated that her team had worked extensively with Alias Wavefront and asked whether SoftImage was compatible with it. If so, this could save time and effort. J-DF replied that although Alias Wavefront was better known than SoftImage, it had no real-time capability and could not support frame-based applications. Bruno Arnaldi (INRIA, Rennes) thought it unlikely that the two systems could interact. PJW stressed that blanket standardization was not necessary as long as the subgroups could work together satisfactorily. It was agreed to test SoftImage to see for which applications it could work and to share existing knowledge re. the strong points of each package.

    The possibility of licensing all sites to test SoftImage was envisaged, especially since F, B and CH sites could obtain it on a free-trial basis for up to 1 year. For other sites price could be negotiated. PJW asked J-DF to check exact terms for multi-site trials asap, so that a firm decision could be arrived at by October at the latest. Could Wavefront be tested in a similar way? N. Thalmann's team could provide test results for W/front in their chosen tasks.
    § J-DF to report on terms

    It was agree that all sites should try running their respective work on all available packages. D. Thalmann proposed using shared memory only, to avoid the need for an interface and potential convention problems. Own data structures should be kept. A module to access shared memory needed to be developed.

    Gerard Hegron (EMN) felt that problems might well arise for ten sites trying to work together, but were unlikely to hinder the efforts of the average sub-group of 2/3 sites. J-DF felt that for the purposes of most sub-groups (not Group 2) it would be easier to use Wavefront, but that ultimately the project would have to develop a common platform.

    The issue of Wavefront's life expectancy was raised, since its successor, Maya, was expected to be released in the near future. N. Thalmann knew of another inexpensive and increasingly common package: "3D studio". Its real-time and 2D capabilities would have to be verified.

    It was agreed that each site would experiment with one or the other of the three packages thus identified (SoftImage - 4/5 sites, Wavefront - 3/4 sites, 3D studio - 3/4 sites) and would report back in time for October meeting.
    § NMT to check on special deal for group use

    The discussion then focused on what to test and it was agreed to try and establish a checklist of functionalities for testing. [This was done in the afternoon, after the close of the formal meeting, and the list can be found in Annex 1]. WP coordinators were to ensure that their WP needs were met. It was further suggested that since the project would be running for four years, developments in other packages (eg. VRML 2.0) should be monitored. However, PJW warned against the risk of turning the project into a software evaluation exercise.

  6. Technical meetings
    Their aim is to bring together the actual researchers working on the various tasks. None had taken place yet, and WP coordinators were asked to organize one for their sub-group within the next 2-3 months. This was agreed, although no tentative plans had yet been made. Participation in relevant conferences was desirable but was no substitute for visits, since mobility was paramount. PhD and post.docs should be encouraged to travel and expenses incurred in this connection could certainly be met from project funds.
    § Coordinators: organize technical meeting

  7. Date of next meeting
    Gerard Hegron (Ecole des Mines, Nantes) confirmed his willingness to host the next Progress meeting which was agreed to take place on Friday, October 3. Participants were invited to travel out on Thursday so as to ensure optimum use of time on Friday. UIB volunteered to host the April meeting. Exact date to be agreed.

  8. Any other business
    John Patterson informed the meeting that an example of work undertaken under the project could be found at http://www.turing.gla.ac.uk.

    Forthcoming conferences: In addition to CA'97, opening in Geneva the next day, the following were mentioned:
    SIGGRAPH 2-8 August, Los Angeles
    Workshop on Animation & Simulation 2-3 September, Budapest
    Eurographics 4-8 September, Budapest
    Interactive Telecom VRST, 15-18 September, Lausanne
    VSMM, 10-12 September, Lausanne

The formal meeting closed at 12 noon, but the whole group gathered in the afternoon to draw up the list of requirements against which to validate the proposed packages. The outcome is appended below, where E=essential and D=desirable.

REQUIREMENT WP2 WP3 WP4 WP5 WP6 WP7 WP8
1. Direct access to the data structure E E E E E   E
2. Independent plug-ins E E E E     E
3. User data linked to the elements of the data structure E E E E E   E
4. Dynamic modification of the data structure E E E E E   E
5. Notify the application of modifications to the geometric, animation and user interface data structures E E D E E   E
6. Support for surface-based rigid bodies and articulated figures     D E      
7. Generic mechanism for collision detection   E D E      
8. General picking mechanism   E E       E
9. User definition of data types E D E E     E
10. Input drivers and mechanisms (real-time VR peripherals) E E D        
11. Output drivers for stereo   E D        
12. User-expandable file formats     E E     E
13. Real-time visualisation E E E D     E
14. Support for key-framing   D E E     E
15. Extendable scripting language   D E E     E
16. Multi-threaded (incl. autonomous processes) D E D D E   E
17. File formats supported E E E E E E E
18. Communication between plug-ins D   D E     D
19. Process communication (esp. across a network) D E D D E   E
20. Identify platforms supported     D     E