-
Notifications
You must be signed in to change notification settings - Fork 0
/
samuellempa.txt
175 lines (174 loc) · 10.3 KB
/
samuellempa.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
* damons is now known as damons-away
* damons-away is now known as damons
* samuell ([email protected]) has joined
<anniepoo> hey samuell!
<anniepoo> 8cD
<samuell> Hi anniepoo !
<anniepoo> how's life?
<samuell> It's good, only been a little hectic lately :)
<samuell> But improving now
<anniepoo> oh good
<anniepoo> 8cD I'm messing about a bit with a tangible interface
<samuell> Aha
<samuell> How does a TUI for prolog look like? :D
<anniepoo> at the moment, like hold up fiducials in front of camera and have Prolog spit out data
<anniepoo> I was thinking that often in the Prolog world we're building UI's for projects with a limited scope
<anniepoo> because Prolog is used so much in research
<anniepoo> and once you get over the basic hurdle of making data come in, it actually reduces the work of building a UI
* freeman42y (~freeman42@2a02:8084:e82:4880:d9b0:a562:2678:f0c) has joined
<samuell> Sounds like interesting ideas
<anniepoo> 8cD
<samuell> Prolog definitely needs to be connected more to all the "real world" stuff
<anniepoo> suppose you're working on something graph like
<anniepoo> a TUI is a great tool for working with a graph
<samuell> Gotcha
<anniepoo> the one really commercially successful TUI is the Reactable
<anniepoo> which is used to set up flow programming in real time to make music
* freeman42x has quit (Ping timeout: 240 seconds)
<samuell> googling now ... looks interesting ... I'm quite a fan of flow-based programming too :)
<anniepoo> me too
<samuell> A bit inpired by Bret Victor's "Seeing Spaces" video perhaps?
<samuell> The reactable
<anniepoo> hmm
<anniepoo> I don't know that video
<samuell> It's a must watch, I think :) http://worrydream.com/SeeingSpaces/
<yesbot> Title: Seeing Spaces (0.06s)
<anniepoo> watching now!
<anniepoo> you know, you asked a good question earlier
<anniepoo> <samuell> How does a TUI for prolog look like? :D
<samuell> :)
<anniepoo> the "for prolog" is something I hadn't really asked in a serious way
<anniepoo> can I use a TUI to aid in writing Prolog?
<samuell> Good question ...
<samuell> I mean, normal predicates goes closely towards the graph thinking
<anniepoo> SWISH's graphic output is a very cool thing. We're using it a lot at work. Make queries in 'fiddle at repl' mode and get graphs as answers
<samuell> ... but for more complex rules ... I don't know
<samuell> Right
<anniepoo> ok, how is a predicate a graph?
<anniepoo> (predicates)
<samuell> Yea, rules always produce concrete results (graphs), so that makes sense
<samuell> But how to represent graph-producing rules? ... in a way more like templates for graphs, right?
<anniepoo> I used to work in Second Life
<anniepoo> SL is a great environment to explore data
<anniepoo> in fact, I built a tool to do so under contract with Univ. of Mich
<anniepoo> you fill in an excel spreadsheet and connect it with this tool in second life
<samuell> Aha
<anniepoo> it makes objects plot the data
<samuell> New thing to me. I was much into 3D stuff some 10-15 years ago, but then got buried in more technical stuff :)
<samuell> Should take a look
<anniepoo> original application was to arrange images of cultures
<anniepoo> so you could have the images arrange by temp, length of growth, conc. of chemical
<anniepoo> It's quite useful for displaying graphs
<anniepoo> because 3D expands more rapidly than 2D
<samuell> interesting
<anniepoo> and in SL you can fly around inside the graph, exploring it
<samuell> Re prolog queries ... they say SQL queries are actually also expanded to flow-based programs behind the scenes ...
<samuell> ... which is graphs ... only grpahs of processes rather than of data.
* TheCommieDuck ([email protected]) has joined
<anniepoo> hmm
<anniepoo> oh, maybe a more real time debugger
<samuell> ah, yea, perhaps
<anniepoo> almost like watching someone work math on a whiteboard
<anniepoo> or else you 'wipe' with a tool, and it does work
<samuell> the idea i found interesting is about being able to get a both intuitive and precis understanding of a complex system
<samuell> or a simple one for that matter
<anniepoo> yes
* stis_ has quit (Ping timeout: 255 seconds)
<anniepoo> looking at seeingspaces poster - the part about NASA control room
<anniepoo> the problem with something like a CAVE is that it's so expensive it's not something you can just use without thinking
<anniepoo> Something we've never done, but I think about doing, is putting a bunch of computers around the shop, so we can, for example, display information rathern than have plans
<anniepoo> right now we have paper plans
<anniepoo> previous worker had a tablet - I made PDF of the plans, and he built from that
<anniepoo> but the paper's interactive - we make notes on the plans as we build
<anniepoo> and in an ideal world, there'd be a projector over the work table
<anniepoo> and you could call up the drawing and use it as a template to cut the part out
<anniepoo> in fact we do a bit of this - we laser cut templates from cardboard or MDF
<anniepoo> to cut sheet metal from
<samuell> ah
<anniepoo> not to mention all the 3D printing/computer controlled machinery/LOM we do
<samuell> heh, i think the paper/screen dichotomy is really an unsolved problem ...
<samuell> i can't even hardly decide whether to take notes mostly on paper, or in my markdown textfile based solution on the computer
<anniepoo> in the center of our shop is a large table
<anniepoo> it's covered with the same melamine coating on cheap office worktables
<anniepoo> you can write on it- which has been amazingly useful
* georgios1 has quit (Quit: WeeChat 1.9)
<anniepoo> making involves writing down dimensions
<anniepoo> making small drawings
<anniepoo> feeling free to write on the table is bizarrely useful
<samuell> i can see that :)
<anniepoo> we were mixing up this urethane casting compound
<samuell> but then, if you want to have it into some more exact computerized format (svg or something) ...?
<anniepoo> right - same downsides as usual
* ianandrich has quit (Remote host closed the connection)
<anniepoo> our worker's not very good with math
<anniepoo> mixing the urethane has to be done quickly
<anniepoo> and requires simple math calculations
<anniepoo> I made him a form, he follows the form and writes in numbers
<anniepoo> it's not very organic
<anniepoo> imagine if the scale were connected to computer system,
<anniepoo> and we had a projector
<anniepoo> draw circles where to put part A, part B, cup C, etc. down
<anniepoo> then it follows along as he does the steps
<samuell> projector based augmented reality ... i like it :)
<anniepoo> I'm drawn back to your question
<anniepoo> yes, I have a small projector
<anniepoo> you know, I'm going to start not with the traditional TUI undertable arrangement
<anniepoo> but with the projector over our white table
<anniepoo> and I'm going to put some conduit up there so I can hang things
<anniepoo> and have a dedicated computer that's always around
<anniepoo> I have an old tablet computer
<anniepoo> the 'tablet notebook' that predated 'tablets'
<anniepoo> and I'll mount a kinect and a normal camera overhead
<anniepoo> SeeingSpaces is an interesting idea
<anniepoo> So what would a more organic interaction with Prolog look like?
<samuell> i guess the problem, when going more generic (as opposed to TUIs for specific use cases), is that any physical representation of data has to be soo generic too ...
<samuell> graphs, as you mentioned, being one good example of generic structure
<anniepoo> though in a TUI making your own fiducials is easy
<anniepoo> maybe we can make generalized TUI's for 'shapes' of data
<anniepoo> then a grid can be a city planning tool or an exploration of constraints in some 2D system or whatever
<samuell> yea, one could always have both the generic graph-based one for "generic" data (data for which there aren't specialized TUI solutions yet), and then develop specialized TUIs as we go
<samuell> representing data as graphs is natural i think
<anniepoo> for much data, yes
<samuell> but, finding a good way to represent rules with graphs, might take more thinking
<anniepoo> yes
<anniepoo> it's more interesting to think about representing rules
<anniepoo> rules act as agents in some ways
<anniepoo> eg, you could represent a forward chaining expert system as a bunch of critters
<anniepoo> and put some facts into the system
<anniepoo> and they'd metabolize them and make new facts
<samuell> yea, the data flow between rules is natural to think of as a graph, true
<samuell> i wish swipl had fwd-chaining ... or perhaps it does?
<anniepoo> CHR sorta is
<anniepoo> 3 mins, brb
<samuell> aha
<anniepoo> back
* samuell was reading https://dtai.cs.kuleuven.be/CHR/about.shtml gotta check this more later.
<yesbot> Title: CHR - Constraint Handling Rules (1.09s)
<anniepoo> the best thing out there is Tom Schrijver's powerpoint presso
<anniepoo> it's online somewhere, if you don't find it email me, I'll send it
<samuell> thanks!
<anniepoo> seeingspaces is in service of understanding more of the science behind the physical manifestation
<anniepoo> we were running a forge, using lots of propane gas
<anniepoo> the pressure kept dropping as the gas expanded and cooled
<anniepoo> I got a big tub, put the propane bottle in it and filled it with water
<anniepoo> of course the thermal mass of the water stabilized the temp of the propane
* whoman eyes open wider
<anniepoo> 8cD hey whoman - always something going on on ##prolog!
<anniepoo> how could constraints be represented?
<anniepoo> many of the most powerful concepts of Prolog involve constraint systems
<whoman> yep !! hope everyone is having a good weekend ! we are indeed beginning to see the construction of reality as well , very exciting times
<anniepoo> yes, we're constructing reality
<anniepoo> I was constructing some boards to mount pictures on earlier
<anniepoo> and will be constructing motorcycle jackstands in a bit
<anniepoo> how could we explore an RDF graph?
<samuell> It is interesting that constraints is what is needed to give anything any form at all ...
<anniepoo> not just display, but interact.
<anniepoo> yes, that's true!
<whoman> im working on understanding the great book =)
<anniepoo> clocksin and mellish?
<anniepoo> or ROK?
* samuell needs to leave soon. getting late in sweden
<anniepoo> wonderful conversation, samuell!
<samuell> anniepoo, the same! thanks for the many pointers and ideas
<whoman> thebook*
<anniepoo> LOL - yes, you've moved my thinking about this a lot!