[news.announce.conferences] 2nd International Workshop on Real-time Ada Issues

andy@minster.york.ac.uk (Andy Wellings) (11/10/87)

- -
                     2nd INTERNATIONAL WORKSHOP
                      ON REAL TIME ADA ISSUES
                  Devon, England, 1 - 3 June 1988
        (Sponsored by Ada UK in cooperation with ACM SIGAda)

                        ORGANIZING COMMITTEE

            Angel Alvarez   Ted Baker       John Barnes
          Mark Dowson     Gerry Fisher    Anthony Gargaro
                           Andy Wellings

   The first International  Workshop  on  Real  Time  Ada  Issues
   identified  a number of critical issues that impact on the use
   of Ada in embedded systems, particularly in  `hard'  real-time
   applications.  Interim  recommendations  for  addressing these
   issues appear in the Workshop proceedings  (ACM  Ada  Letters,
   October 1987).

   The 2nd Workshop will examine the critical issues  in  greater
   depth. It is expected that the participants will have specific
   proposals which can be analyzed during  the  workshop  with  a
   view  to  agreeing  recommended  courses  of action to various
   segments of the Ada community. The  issues  to  be  considered
   include:

   SCHEDULING MECHANISMS

   Hard real-time systems may combine periodic transactions which
   must  meet  regular  deadlines with sporadic transactions with
   defined worst-case response times. Such deadline scheduling is
   difficult  to program with Ada's existing  timing and priority
   mechanisms.

        What mechanisms are needed for satisfactory handling
        of   deadlines   for   both  periodic  and  sporadic
        transactions?

        What new  or  modified  Ada  language  features  are
        needed to implement these mechanisms?

   ASYNCHRONOUS TRANSFER OF CONTROL

   Rapid event driven mode shifts  and  certain  kinds  of  error
   recovery  require  a  task  asynchronously  to affect the flow
   control of another task in ways that  cannot  be  conveniently
   achieved  by  abort  alone.  Polling  strategies  reduce  task
   readability and do not, in general, guarantee fast response.

        Are  additional  language  constructs,  such  as   a
        FAILURE exception, needed?

        How  can  these  constructs  be  defined   so   that
        anomalous  cases  eg  when  interacting  with  other
        exceptions, are minimized?


   DISTRIBUTED EXECUTION: RE-CONFIGURATION AND RECOVERY

   Ada, in common  with  most  programming  languages,  does  not
   directly  address  the  problems  of distributed execution. In
   particular, Ada provides no mechanisms to support dynamic  re-
   configuration  of  a  distributed  system  in  case of partial
   hardware failure.

        Should re-configuration be achieved transparently by
        the  run-time  system,  by  a  separate mechanism eg
        system generation and re-generation,  or  under  the
        control of the application program?

        What additional language  constructs,  if  any,  are
        needed?

   The workshop, which will be held at  the  Manor  House  Hotel,
   Moretonhampstead,  Devon,  UK,  will  consist  of  three days'
   intensive consideration  of  these  issues  by,  at  most,  35
   participants. Prospective participants should submit a maximum
   3  page  position  paper  by  14  December  1987,   explicitly
   addressing  one  of  the  workshop  issues  and  suitable  for
   publication in the proceedings. A small number of participants
   will  be  requested  to prepare short keynote presentations to
   initiate discussion.  Position papers (6 copies or  electronic
   mail in UNIX troff form if possible) should be sent to:

                           Andy Wellings
             Computer Science Dept, University of York,
                 Heslington, York YO1 5DD ENGLAND.
                    Tel: +44 904 430000 ext 305
                   Email: andy@minster.york.ac.uk

                                 or

                       (Electronic mail only)
                          Anthony Gargaro
    Email: Gargaro@ada20.isi.edu (or Gargaro@ajpo.sei.cmu.edu)

       Submission of papers by electronic mail is encouraged
- -

andy@minister.york.ac.uk (Andy Wellings) (11/17/87)

- - -
[Note that a position paper is required to attend this workshop and must
 be submitted by December 14th, 1987 to the address listed herein.]
- - -

                     2nd INTERNATIONAL WORKSHOP
                      ON REAL TIME ADA ISSUES
                  Devon, England, 1 - 3 June 1988
        (Sponsored by Ada UK in cooperation with ACM SIGAda)

                        ORGANIZING COMMITTEE

            Angel Alvarez   Ted Baker       John Barnes
          Mark Dowson     Gerry Fisher    Anthony Gargaro
                           Andy Wellings

   The first International  Workshop  on  Real  Time  Ada  Issues
   identified  a number of critical issues that impact on the use
   of Ada in embedded systems, particularly in  `hard'  real-time
   applications.  Interim  recommendations  for  addressing these
   issues appear in the Workshop proceedings  (ACM  Ada  Letters,
   October 1987).

   The 2nd Workshop will examine the critical issues  in  greater
   depth. It is expected that the participants will have specific
   proposals which can be analyzed during  the  workshop  with  a
   view  to  agreeing  recommended  courses  of action to various
   segments of the Ada community. The  issues  to  be  considered
   include:

   SCHEDULING MECHANISMS

   Hard real-time systems may combine periodic transactions which
   must  meet  regular  deadlines with sporadic transactions with
   defined worst-case response times. Such deadline scheduling is
   difficult  to program with Ada's existing  timing and priority
   mechanisms.

        What mechanisms are needed for satisfactory handling
        of   deadlines   for   both  periodic  and  sporadic
        transactions?

        What new  or  modified  Ada  language  features  are
        needed to implement these mechanisms?

   ASYNCHRONOUS TRANSFER OF CONTROL

   Rapid event driven mode shifts  and  certain  kinds  of  error
   recovery  require  a  task  asynchronously  to affect the flow
   control of another task in ways that  cannot  be  conveniently
   achieved  by  abort  alone.  Polling  strategies  reduce  task
   readability and do not, in general, guarantee fast response.

        Are  additional  language  constructs,  such  as   a
        FAILURE exception, needed?

        How  can  these  constructs  be  defined   so   that
        anomalous  cases  eg  when  interacting  with  other
        exceptions, are minimized?


   DISTRIBUTED EXECUTION: RE-CONFIGURATION AND RECOVERY

   Ada, in common  with  most  programming  languages,  does  not
   directly  address  the  problems  of distributed execution. In
   particular, Ada provides no mechanisms to support dynamic  re-
   configuration  of  a  distributed  system  in  case of partial
   hardware failure.

        Should re-configuration be achieved transparently by
        the  run-time  system,  by  a  separate mechanism eg
        system generation and re-generation,  or  under  the
        control of the application program?

        What additional language  constructs,  if  any,  are
        needed?

   The workshop, which will be held at  the  Manor  House  Hotel,
   Moretonhampstead,  Devon,  UK,  will  consist  of  three days'
   intensive consideration  of  these  issues  by,  at  most,  35
   participants. Prospective participants should submit a maximum
   3  page  position  paper  by  14  December  1987,   explicitly
   addressing  one  of  the  workshop  issues  and  suitable  for
   publication in the proceedings. A small number of participants
   will  be  requested  to prepare short keynote presentations to
   initiate discussion.  Position papers (6 copies or  electronic
   mail in UNIX troff form if possible) should be sent to:

                           Andy Wellings
             Computer Science Dept, University of York,
                 Heslington, York YO1 5DD ENGLAND.
                    Tel: +44 904 430000 ext 305
                   Email: andy@minister.york.ac.uk

                                 or

                       (Electronic mail only)
                          Anthony Gargaro
    Email: Gargaro@ada20.isi.edu (or Gargaro@ajpo.sei.cmu.edu)

       Submission of papers by electronic mail is encouraged
- - -