Our first meeting at the Edgar Wallace...
12 views
Skip to first unread message
crishog
Nov 24, 2008, 4:24:53 AM11/24/08
to BAA London
was attended by:
Paul Grosvenor
Stephen Taylor
Ellis Morgan
Chris Hogan
Dick Bowman
Mike Hughes
Chris Paul
John Jacob
Ray Cannon
Phil Last
and we had apologies from Graeme Robertson, John Scholes & Phil
Chastney.
Jake too some minutes & I'll post them this morning
Chris
Paul Grosvenor
Stephen Taylor
Ellis Morgan
Chris Hogan
Dick Bowman
Mike Hughes
Chris Paul
John Jacob
Ray Cannon
Phil Last
and we had apologies from Graeme Robertson, John Scholes & Phil
Chastney.
Jake too some minutes & I'll post them this morning
Chris
Phil Last
Nov 24, 2008, 5:03:26 AM11/24/08
to BAA London
On Nov 24, 9:24 am, crishog <goo...@4xtra.com> wrote:
> was attended by:
>
> Paul Grosvenor
> Stephen Taylor
> Ellis Morgan
> Chris Hogan
> Dick Bowman
> Mike Hughes
> Chris Paul
> John Jacob
> Ray Cannon
> Phil Last
>
> and we had apologies from Graeme Robertson, John Scholes & Phil
> Chastney.
and Kai Jäger> was attended by:
>
> Paul Grosvenor
> Stephen Taylor
> Ellis Morgan
> Chris Hogan
> Dick Bowman
> Mike Hughes
> Chris Paul
> John Jacob
> Ray Cannon
> Phil Last
>
> and we had apologies from Graeme Robertson, John Scholes & Phil
> Chastney.
crishog
Nov 24, 2008, 5:29:49 AM11/24/08
to BAA London
> and Kai Jäger
Oops! Kai was the first to say he'd like to come, but couldn't - my
apologies
Chris
crishog
Nov 25, 2008, 5:34:16 AM11/25/08
to BAA London
The attendees drifted in over a period of time - our fault (C, J & P)
managing to give out 3 different start times for the meeting, but in a
way this helped to establish the social atmosphere of the gathering.
Having no rigid agenda, initially we were split into three groups:
Ellis and Ray; Paul, Stephen & Phil: and Chris, Mike, Jake, Chris &
Dick. This summary is based on some notes Jake took, more detailed
minutes by Ellis and my own (imperfect) memory:
Chris, Mike, Jake, Chris & Dick: We started off from dick's original
suggestion, to discuss APL software libraries,but we soon were
directed onto two other related issues. Mike pointed out a bigger
problem that an absence of code libraries or examples was a lack of
_concept_ examples. Chris Paul also has an immediate education issue
which is rlevant His current project has a new graduate developer &
when told he was to work with an PAL system, he simply went again
"googled" it - a fairly standard reaction from developers in all
fields nowadays - but the first material he discovered was an APL
Joke, which is hardly going to allow him to write or support APL code.
We decided to draw up a list of strategies and tactics we could all
use to help achieve the aims of educating newcomers to APL, convincing
business to use APL and supporting the existing APL community. Grand
ideas and I think they willl be much debated, so I've set the lists up
in the APL Wiki at http://www.aplteam2.com/aplwiki/APLMeetingAtTheEdgarWallace
were I think it will be easier to produce a consolidated document,
rather than in a google discussion group.
After some time I called us all together for a general meeting at the
end. Amongst other topics that of encouraging new people to take up
APL came up. My view remains that the business of ensuring the future
of APL does not rest with a dozen old and nearly old men in a pub in
London, however it is a valid BAA aim and we can help
On helping others to learn APL. I have sent copies of the "learnAPL"
poster in the recent Vector to a couple of friends whose children are
OT oriented and approaching university age.
You may not necessarily want to find new APLers from that group of
sad souls who enjoy a Sudoku puzzle, but do you think there would be
any interest in an "BAA APL Sudoku open source project"? I would
happily donate a bit of my time to this and give the BAA the copyright
to my code. May be others would do the same. There is a health
warning: my design criteria have changed many times but no code has
been thrown away; new features have been added leading to overuse of
data fields; there is no real documentation for users or coders; there
are no (well very few) functions I would be happy to expose to
seasoned APL coders, let alone beginners; ..
We briefly debated the need for a name for our group (APL unplugged or
APL rebooted) - Dick rightly said he didn't care what the name was as
long as we actually did what we said we would do. With Paul & Stephen
present (both BAA committee members) & everyone else being a BAA
member we decided that, rather than an informal event organised by
Chrisl, Jake & Phil, this was now the London chapter of the British
APL Association. While not formally agreeing a timetable of meetings,
it was suggested that we use the planned meeting of ex-IP Sharp
employees as a Christmas venue, provisionally named "BAA-Humbug". The
main purpose I can see of a name is a simple "tag" something we all
refer to in such a way that it spreads (via word of mouth or google
searches), again to assist in raising our profile.
Ellis & Ray: Ray has a Pocket APL Sudoku workspace that also runs
under windows. It is designed with the small pocket display in mind
uses the "wand" and no keyboard. Ellis's is written under Dyalog APL
10.1 and seems to run OK under version 12. It runs almost entirely in
the session using keyboard input with little need for GUI or mouse.
Both of them have tried to write code that follows and evaluates the
way a human solves puzzles rather than to optimise a computer
solution. Ray then went on to show Ellis his GPS code which has two
threads running, the first monitors his GPS devices and issues windows
events based on the readings, the second catches these events and
interprets them on maps and other displays for the user.
Ellis also spoke to Stephen about the difficulty he has have in
entering APL characters into the AplWiki editor direct from the
keyboard. Stephen was able to demonstrate one possible cause for my
difficulty and Ellis is now seeking a solution in the DyalogUsers
group.
Paul, Stephen & Phil: Sorry lads, no minutes from you, could someone
fill us in?
For any errors in this document, my apologies in advance. Can everyone
have a look at the strategies and tactics lists on the Wiki & feel
free to add to or amend the list if you remember something I've
forgotten. I'm aware that I said I'd get this out & it's Tuesday
already.
Chris
managing to give out 3 different start times for the meeting, but in a
way this helped to establish the social atmosphere of the gathering.
Having no rigid agenda, initially we were split into three groups:
Ellis and Ray; Paul, Stephen & Phil: and Chris, Mike, Jake, Chris &
Dick. This summary is based on some notes Jake took, more detailed
minutes by Ellis and my own (imperfect) memory:
Chris, Mike, Jake, Chris & Dick: We started off from dick's original
suggestion, to discuss APL software libraries,but we soon were
directed onto two other related issues. Mike pointed out a bigger
problem that an absence of code libraries or examples was a lack of
_concept_ examples. Chris Paul also has an immediate education issue
which is rlevant His current project has a new graduate developer &
when told he was to work with an PAL system, he simply went again
"googled" it - a fairly standard reaction from developers in all
fields nowadays - but the first material he discovered was an APL
Joke, which is hardly going to allow him to write or support APL code.
We decided to draw up a list of strategies and tactics we could all
use to help achieve the aims of educating newcomers to APL, convincing
business to use APL and supporting the existing APL community. Grand
ideas and I think they willl be much debated, so I've set the lists up
in the APL Wiki at http://www.aplteam2.com/aplwiki/APLMeetingAtTheEdgarWallace
were I think it will be easier to produce a consolidated document,
rather than in a google discussion group.
After some time I called us all together for a general meeting at the
end. Amongst other topics that of encouraging new people to take up
APL came up. My view remains that the business of ensuring the future
of APL does not rest with a dozen old and nearly old men in a pub in
London, however it is a valid BAA aim and we can help
On helping others to learn APL. I have sent copies of the "learnAPL"
poster in the recent Vector to a couple of friends whose children are
OT oriented and approaching university age.
You may not necessarily want to find new APLers from that group of
sad souls who enjoy a Sudoku puzzle, but do you think there would be
any interest in an "BAA APL Sudoku open source project"? I would
happily donate a bit of my time to this and give the BAA the copyright
to my code. May be others would do the same. There is a health
warning: my design criteria have changed many times but no code has
been thrown away; new features have been added leading to overuse of
data fields; there is no real documentation for users or coders; there
are no (well very few) functions I would be happy to expose to
seasoned APL coders, let alone beginners; ..
We briefly debated the need for a name for our group (APL unplugged or
APL rebooted) - Dick rightly said he didn't care what the name was as
long as we actually did what we said we would do. With Paul & Stephen
present (both BAA committee members) & everyone else being a BAA
member we decided that, rather than an informal event organised by
Chrisl, Jake & Phil, this was now the London chapter of the British
APL Association. While not formally agreeing a timetable of meetings,
it was suggested that we use the planned meeting of ex-IP Sharp
employees as a Christmas venue, provisionally named "BAA-Humbug". The
main purpose I can see of a name is a simple "tag" something we all
refer to in such a way that it spreads (via word of mouth or google
searches), again to assist in raising our profile.
Ellis & Ray: Ray has a Pocket APL Sudoku workspace that also runs
under windows. It is designed with the small pocket display in mind
uses the "wand" and no keyboard. Ellis's is written under Dyalog APL
10.1 and seems to run OK under version 12. It runs almost entirely in
the session using keyboard input with little need for GUI or mouse.
Both of them have tried to write code that follows and evaluates the
way a human solves puzzles rather than to optimise a computer
solution. Ray then went on to show Ellis his GPS code which has two
threads running, the first monitors his GPS devices and issues windows
events based on the readings, the second catches these events and
interprets them on maps and other displays for the user.
Ellis also spoke to Stephen about the difficulty he has have in
entering APL characters into the AplWiki editor direct from the
keyboard. Stephen was able to demonstrate one possible cause for my
difficulty and Ellis is now seeking a solution in the DyalogUsers
group.
Paul, Stephen & Phil: Sorry lads, no minutes from you, could someone
fill us in?
For any errors in this document, my apologies in advance. Can everyone
have a look at the strategies and tactics lists on the Wiki & feel
free to add to or amend the list if you remember something I've
forgotten. I'm aware that I said I'd get this out & it's Tuesday
already.
Chris
Dick Bowman
Nov 25, 2008, 8:49:08 AM11/25/08
to baa-l...@googlegroups.com
crishog wrote:
> [... deleted ...] everyone else being a BAA
> member [... deleted ...]
The above not quite true.
> [... deleted ...] everyone else being a BAA
> member [... deleted ...]
The above not quite true.
Anyway - I think it would be good for us to actually achieve something
before next meeting.
Can I suggest that by the end of the week we each commit to doing "one
little thing" - I'm a great believer in the five-minute project, doesn't
take much out of the day, but keep on doing it...
That's my five minutes done for today.
Stephen Taylor <editor@vector.org.uk>
Nov 27, 2008, 1:17:00 AM11/27/08
to BAA London
Phil & Paul spent some time looking at techniques I developed for a
new internal system at Dyalog, written from scratch in Dyalog 12
Unicode. Some features:
–all persistent data held in an SQL Server database, even enumerators;
most tables have single, sequential numeric keys; enumerators also
have a Name column
–all source code SALted and assembled in the ws when the application
starts; the appn can also be run from a ws saved after 'assembling' it
from the scripts
GUI WRITTEN ENTIRELY USING []NEW:
–a SmartForm class that adds layout tools to the native Dyalog GUI
Form object
–a group of classes that extend other native GUI controls: SmartEdit,
SmartCombo, etc all support Caption, Align, Value and Missing
properties; these controls all manage their own labels, which move,
dim or hide with the control
–an AnchoredCombo class, that relates a SmartCombo to a database
table: unless Missing is true, Value is always a member of the table
key, Text the corresponding Name
–an AnchoredForm class with methods such as Load, Serialise and Edit
to handle editing a row from main data tables
–a group of 'business object' classes derived from AnchoredForm to
edit specific tables, eg Company, Contract, Licence
I demonstrated how in the session I could type
q <- []NEW Company 82
q.Edit
and the form would appear for company 82, ready to edit. Because the
AnchoredForm methods (Load and Serialise) map between table rows and
GUI controls, the methods of 'business object' classes such as
Company, Contract and Licence read very clearly. Eg
{disclose} CompanyID.Missing {rotate} (ContractID
CompanyID).Value
The application user interface is a singleton class, UI, also derived
from SmartForm. It has menu and toolbars but gives most of its screen
space to a Microsoft Web Browser OCX control. UI has a variable
ThisObj. Simple navigation round the app consists of instantiating the
object you want and assigning it to ThisObj, then displaying the
results of ThisObj.Display.
HTML-BASED DISPLAYS
All the 'business object' classes support the Display method,
returning an HTML document for UI to display in the browser OCX
control. HTML/CSS is a richer language for layout than native GUI
controls. Styles and positioning can be adjusted simply by editing the
application's CSS file and refreshing the screen; viewing displays
with Firefox and Firebug is even more helpful.
The displays are thick with hyperlinks to related objects. For
example, the display of a reseller company lists its clients as
hyperlinks to their Company records, such as <a href="javascript:show
Company 82">A. Client</a>
Other View classes, such as Dashboard, Log and Search, also support
the Display method.
SUMMARY
Using []NEW in this way required figuring out a few tricks, (thanks
here to ProfDoc for the tip on Javascript) but produced very good
encapsulation. Burying the object/table and enumerator/text mappings
in a few 'anchored' classes eliminated a lot of tedious code from the
'business object' methods. Objects behave in a very APL-like way: when
they disappear, it's like a function clearing the stack: all the
picnic litter got taken home. The whole app runs inside an anonymous
instance:
([]NEW UI).Wait
When Wait clears the stack nothing remains in the root all but the
original defined classes and namespaces.
I used a subset of these techniques to write the single-script
implementation of Solitaire on the APL Wiki.
SJT
new internal system at Dyalog, written from scratch in Dyalog 12
Unicode. Some features:
–all persistent data held in an SQL Server database, even enumerators;
most tables have single, sequential numeric keys; enumerators also
have a Name column
–all source code SALted and assembled in the ws when the application
starts; the appn can also be run from a ws saved after 'assembling' it
from the scripts
GUI WRITTEN ENTIRELY USING []NEW:
–a SmartForm class that adds layout tools to the native Dyalog GUI
Form object
–a group of classes that extend other native GUI controls: SmartEdit,
SmartCombo, etc all support Caption, Align, Value and Missing
properties; these controls all manage their own labels, which move,
dim or hide with the control
–an AnchoredCombo class, that relates a SmartCombo to a database
table: unless Missing is true, Value is always a member of the table
key, Text the corresponding Name
–an AnchoredForm class with methods such as Load, Serialise and Edit
to handle editing a row from main data tables
–a group of 'business object' classes derived from AnchoredForm to
edit specific tables, eg Company, Contract, Licence
I demonstrated how in the session I could type
q <- []NEW Company 82
q.Edit
and the form would appear for company 82, ready to edit. Because the
AnchoredForm methods (Load and Serialise) map between table rows and
GUI controls, the methods of 'business object' classes such as
Company, Contract and Licence read very clearly. Eg
{disclose} CompanyID.Missing {rotate} (ContractID
CompanyID).Value
The application user interface is a singleton class, UI, also derived
from SmartForm. It has menu and toolbars but gives most of its screen
space to a Microsoft Web Browser OCX control. UI has a variable
ThisObj. Simple navigation round the app consists of instantiating the
object you want and assigning it to ThisObj, then displaying the
results of ThisObj.Display.
HTML-BASED DISPLAYS
All the 'business object' classes support the Display method,
returning an HTML document for UI to display in the browser OCX
control. HTML/CSS is a richer language for layout than native GUI
controls. Styles and positioning can be adjusted simply by editing the
application's CSS file and refreshing the screen; viewing displays
with Firefox and Firebug is even more helpful.
The displays are thick with hyperlinks to related objects. For
example, the display of a reseller company lists its clients as
hyperlinks to their Company records, such as <a href="javascript:show
Company 82">A. Client</a>
Other View classes, such as Dashboard, Log and Search, also support
the Display method.
SUMMARY
Using []NEW in this way required figuring out a few tricks, (thanks
here to ProfDoc for the tip on Javascript) but produced very good
encapsulation. Burying the object/table and enumerator/text mappings
in a few 'anchored' classes eliminated a lot of tedious code from the
'business object' methods. Objects behave in a very APL-like way: when
they disappear, it's like a function clearing the stack: all the
picnic litter got taken home. The whole app runs inside an anonymous
instance:
([]NEW UI).Wait
When Wait clears the stack nothing remains in the root all but the
original defined classes and namespaces.
I used a subset of these techniques to write the single-script
implementation of Solitaire on the APL Wiki.
SJT
Stephen Taylor <editor@vector.org.uk>
Dec 2, 2008, 9:45:59 AM12/2/08
to BAA London
Someone undertook to set this up with Graeme, convenor of the IPSA
Xmas parties. I believe the idea is to start 2-3 hours before the
party. But I don't think I've yet seen a date…
Was that Michael? What's happening?
Stephen
On Nov 25, 10:34 am, crishog <goo...@4xtra.com> wrote:
> The attendees drifted in over a period of time - our fault (C, J & P)
> managing to give out 3 different start times for the meeting, but in a
> way this helped to establish the social atmosphere of the gathering.
>
> Having no rigid agenda, initially we were split into three groups:
> Ellis and Ray; Paul, Stephen & Phil: and Chris, Mike, Jake, Chris &
> Dick. This summary is based on some notes Jake took, more detailed
> minutes by Ellis and my own (imperfect) memory:
>
> Chris, Mike, Jake, Chris & Dick: We started off from dick's original
> suggestion, to discuss APL software libraries,but we soon were
> directed onto two other related issues. Mike pointed out a bigger
> problem that an absence of code libraries or examples was a lack of
> _concept_ examples. Chris Paul also has an immediate education issue
> which is rlevant His current project has a new graduate developer &
> when told he was to work with an PAL system, he simply went again
> "googled" it - a fairly standard reaction from developers in all
> fields nowadays - but the first material he discovered was an APL
> Joke, which is hardly going to allow him to write or support APL code.
> We decided to draw up a list of strategies and tactics we could all
> use to help achieve the aims of educating newcomers to APL, convincing
> business to use APL and supporting the existing APL community. Grand
> ideas and I think they willl be much debated, so I've set the lists up
> in the APL Wiki athttp://www.aplteam2.com/aplwiki/APLMeetingAtTheEdgarWallace
Xmas parties. I believe the idea is to start 2-3 hours before the
party. But I don't think I've yet seen a date…
Was that Michael? What's happening?
Stephen
On Nov 25, 10:34 am, crishog <goo...@4xtra.com> wrote:
> The attendees drifted in over a period of time - our fault (C, J & P)
> managing to give out 3 different start times for the meeting, but in a
> way this helped to establish the social atmosphere of the gathering.
>
> Having no rigid agenda, initially we were split into three groups:
> Ellis and Ray; Paul, Stephen & Phil: and Chris, Mike, Jake, Chris &
> Dick. This summary is based on some notes Jake took, more detailed
> minutes by Ellis and my own (imperfect) memory:
>
> Chris, Mike, Jake, Chris & Dick: We started off from dick's original
> suggestion, to discuss APL software libraries,but we soon were
> directed onto two other related issues. Mike pointed out a bigger
> problem that an absence of code libraries or examples was a lack of
> _concept_ examples. Chris Paul also has an immediate education issue
> which is rlevant His current project has a new graduate developer &
> when told he was to work with an PAL system, he simply went again
> "googled" it - a fairly standard reaction from developers in all
> fields nowadays - but the first material he discovered was an APL
> Joke, which is hardly going to allow him to write or support APL code.
> We decided to draw up a list of strategies and tactics we could all
> use to help achieve the aims of educating newcomers to APL, convincing
> business to use APL and supporting the existing APL community. Grand
> ideas and I think they willl be much debated, so I've set the lists up
Phil Last
Oct 8, 2018, 1:00:07 PM10/8/18
to BAA London
I just noticed that next month, November 2018, will be 10 years since we
started at the Edgar Wallace pub in Essex Street. From this, the first post in
this google group I infer that the meeting was held on Friday
2008-11-21
My announcement for this month's meeting assumes not many of us will be here on October 26.
Can everyone think about the next one and make it worthwhile?
-------------------------------------------------------Reply all
Reply to author
Forward
0 new messages