Let's talk meta...

[wyleu]

I've had a coincidence!

On the day I replace the flaky X-10 AD10's, I can type,

python>
>>>import visual
>>>a = visual.sphere()

and it draws a sphere!
Visual python runs on my dell laptop in ubuntu studio at last!

[wyleu]

Mind you it's flat shading all the objects...

[StBarnabas]

Wyleu
Sorry. Just seen this post. Get less chance to look at the forum these days.
i would like to learn python - indeed python is now being used at places like CERN so must be good, but have been using MATLAB and a bit of OCTAVE for some years (indeed MATLAB since the mid '80s!). In Matlab/Octave the command   "sphere" will do something similar...

[wyleu]

It's been a fair old wander that's for sure. If I try to explain what I've been trying to do and why it's not come about then perhaps we might have a basis for discussion.

I've been interested in social measurement networks for years. I've tried to flog the idea to government on a couple of occasions without much success and a lot of paperwork, I've tried to build it myself and with the help of one or two people around here I've done one hell of a lot of hours on it all.

Mostly I see early adoption with all the opinion, expertise, effort, laziness and goodness knows what else? No matter it's all good stuff and really it's all to answer a couple of simple questions. Can I manage my resources? Now we've got a fair old range there but from a measurement perspective it boils down to rates and rythms ( I word I can never spell ).

We're dead lucky we have cheap accurate digital sensors and that's amazing. We've tended to concentrate on one-wire but the problem probably has many answers. I have always wanted to hear from someone in the passenger aircraft industry on the subject, They probably know a fair bit more.

But what are we looking at?

Well I might 'build' a flow sensor consisting of a low pressure flap value with a magnet on it, a ratiomatic hall effect probe and a couple of thermometers above and below the flap, a clever little half a bee thing and a one wire dongle. But what then?
I inevitably sit down with One wire viewer and read them as temperatures, just till I get it working... and that of course never quite happens because in a lot of cases we're well short on reliability of the network. Not inherently one-wire specifically, but dratted PC's and mains bumps and brown outs and so it goes on... The actual since we're trying to control these very things we've got to have reliability above that of the utilities to lend us any credibility at all.

So we build a lot of hardware, which is fantastic! and this seeds ideas about quite what we want to see.

Really we want some level of abstraction above the simple sensors that can actually represent a instance of the thing we'd love to use to monitor, ( Daft flap valvy thing...), so theres a responsibility to provide documentation to describe the beast, and provide both a description of the object itself and where this particular instance ( serial No 1 probably followed by no others) is doing what ever it's meant to do.

Now I think this is all 3D data as I feel the conventional 2D block diagram doesn't convey all the information, so obviously thingy needs a location in space and by implication it wants to link to some form of geographically indexable object like a house. We have effectively a place in space and we can start populating our virtual environment with objects. Obviously I'd like a fully ray traced image of my device in the environment but I'd actually really settle for a lot less. and really the argument is not so much about what the object looks like when it's in the environment but what default object represents it until it's got some metadata to describe itself.

I assume a sphere, perhaps others prefer a cube or a pyramid, it doesn't matter but that is why my fascination with the idea of typing
 a = sphere()
, it's more to do with not having to specify parameters that appeals, and it represents the point at which the environment starts to react. Because by changing a parameter at some stimulation, a difference in temperatures perhaps, we could do tricks like change its colour.

a.red = 255
a.blue = 0
a.green = 0

Suddenly our sphere changes colour to tell us the thingy has detected something interesting.

Now the problems. It would be really good if we had a standard way of describing this, and it wants to be completely agnostic and open and free. It's a big multi user dungeon of an environment and to actually do it right requires all kinds of things that don't work together right now to do so faultlessly. Sadly nearly everyone is a bit proprietary about this stuff and it could quickly descend into the fragmented world of your distro is better than my ipod which is better than your PC etc.

There also needs to be a mapping of units. I'm actually measuring valve flap angle, but the transfer characteristic of flap angle to flow rate is more of an exercise in successive approximation than any hard one equation and needs to be compared with other sensors to really develop any sort of accurate statement. So our environment needs to be able to modify these parameters, easily in some fairly perverse ways. Nice and accurate function generation and table look up plus goodness knows what else. A function to convert temperature to colour. I wrote something that did this sort of thing but it was a bit slow and clunky and taught me a lot about why I shouldn't do it the way I did ! I'll bet mathslab does it far better.

If we could standardize on the description of the environment and the transfer functions we would really have something useful, but I'm caught up on quite how one proceeds. I've had the system rendering images of tanks and such like with temperatures on but it needs a game like authoring environment to define the objects and transfer functions. Sketch up looked pretty good but google didn't seem to like linux for some reason.

[Baz]

It is very difficult to understand what you are trying to do but I think you mean:
"I wish to establish a standard graphical interface for displaying data on a PC relating to home equipment and environment. An example being water flow rate".

Using any specific programming language will immediately create a major constraint. For maximum adoption you should base the display on simple HTML & javascript and not use any features that make it browser specific, not require recent browser versions, and especially not use Flash.

The page code can then be produced by a variety of languages.

Python seems to be growing in popularity even with some professionals owing to the easy access of the editor/interpreter relative to Perl. I too must get round to learning it.

[stephendv]

HTML 5 support for X3D could be useful: http://www.web3d.org/about/overview/

"X3D is a royalty-free open standards file format and run-time architecture to represent and communicate 3D scenes and objects using XML. It is an ISO ratified standard that provides a system for the storage, retrieval and playback of real time graphics content embedded in applications, all within an open architecture to support a wide array of domains and user scenarios."

[wyleu]

HTML 5 support for X3D could be useful: http://www.web3d.org/about/overview/

"X3D is a royalty-free open standards file format and run-time architecture to represent and communicate 3D scenes and objects using XML. It is an ISO ratified standard that provides a system for the storage, retrieval and playback of real time graphics content embedded in applications, all within an open architecture to support a wide array of domains and user scenarios."

Yes that the sort of thing. I tried kicking around various bits of VRML along the way but they all seemed to stop at drawing a Teapot and got no further without some arcane error message. OpenGL would seem to be involved as well but after trying to get various implementations of that to work around browsers was ...involved.
Having spent a little time dealing with various so called standards along the way I'd prefer to see something that actually running rather than the inevitable coming soon. My personal belief is that something is more likely to emerge from the games community who don't really look over their shoulders too much rather than any organisation with other interests to serve.

As to which language gets used why should anyone care? I use python cos I got sick of having to learn something new for every new piece of kit we bought. If you are starting from scratch it's probably as close to a critical mass as you will find, but certainly other opinions exist. It's sad that Apple seem to be linking apps to language to stores but it's understandable, and Java under Oracles tutorage could do something very strange tomorrow. At the end of the day it shouldn't be a stumbing block but if small linux type bricks can get start to provide such tools then the general programming world would seem to indicate that languages of this sort are a useful aid to fast and relatively fault free programming.

In answer to Baz,
it's a bit more than 'just' a graphical interface. That is really only something that realistically would get viewed during setting up, the inevitable look at my new toy phase, and then after a period of long neglect the inevitable fault finding. The transfer function element is probably the level we should be most concerned about. The excellent half-bee initiative is already scratching the edge of this debate with the replacement of the no longer produced 4 channel AtoD replacement and if we don't consider this sort of thing, fairly carefully now, we stand a chance of a great deal of fragmentation.
Only by polling a community such as this can we possibly generate anywhere near a decent subset of functionality. Just think about flow rate, since it's a possible example bench mark. It's a concept that will have a massive range of interpretations across different areas of expertise and althou' the measurements simply boil down to mass passing a point, it is unlikely that what would satisfy a wet solar designer would be particularly relevant to wind turbine bods or river turbine individuals. But that's not to say anything about such virtual concepts that people would be interested in such as efficiency or reliability.

What I would like this thread to be is a clearing house for the existing standards that are out there. Having watched the direct collision between the RGB world of video and CMYK empire of print and tried to line monitors up for both, I'm sure this is more of a collection of information project rather than a crazed session of code writing, but the more people who can set finger to interpreter, with confidence would be of great benefit.

As is obvious this is all a bit wishy washy but it if you have a little time to think about it then please do so and stick your thoughts down. Your interest makes them as relevant as anyone else's because the quality is the idea itself rather than our rather modern obsession with who might or might not have delivered the message.

[KenB]

Python is gaining popularity amongst professionals.

The whole of the Onzo web back-end system was written in Python - we had to go to the ends of the earth (Sweden actually) to find a young Python programmer who would work for "an engineer's salary" and not be hawking his very marketable python skills in the financial sector - and earning a small fortune.

Good work Wyleu, we need all sorts of skills to make these ideas work - and I will be focussing on the lowest cost hardware - because that's my comfort zone, and leave the clever programming stuff to my betters.



Ken

[KenB]

Speaking of meta data you might be interested in a monthly tech meetup held in London.

http://www.meetup.com/minibar/

Minibar is held on the 4th Friday evening of the month from about 7pm and is open to anyone working in the tech field.  Basically they take over a bar for the evening and hand out free beer, whilst encouraging people to talk about their ideas, or pitch their latest start-up idea in just 20 seconds.

It's normally attended by about 275 people - from all over the emerging new tech community.  If you like an occasional night out in London, drinking (free) beer and socialising with fellow geeks and geek-chicks - then this might be just for you.


Ken

[StBarnabas]

Wyleu

great to have more of your input. You have been much missed! I understand exactly where you are coming from but like Ken my main thrust at present is  still on sensors.

Incidentally just back from holiday and 3 of Ken's recommendations for  Chinese flow sensors have arrived.

I do not understand database well but can write algorithms well. If there were a critical mass on the forum I would learn Python etc. As said elsewhere been programming since the early '70s had my first email address in '78  and true 'internet' access in '81 (well managed node 42 of SERCNet).

My experience is that sadly most on the forum are not interested in this and those with the appropriate skills are very busy (As Ken pointed out such skills are rare and highly desired). The problem is that we are a bit desperate. I look at OCTAVE as I have been using MATLAB since the mid '80s though I could go down a 'C'/ ROOT route (CERN) or be persuades to use Java/Swing or indeed PYTHON. My problem is that I have very little time. In MATLAB I can do most of what I want very quickly. This is because of 25 years familiarity - having written 100s of thousands of lines of code and having all my code from about 1990.

Convince me to change!

[EccentricAnomaly]

+1 for Python.

I've tried a few languages of recent and it's the one I feel most comfortable in.

I was doing a lot of professional work in C++ a couple of years ago and it gave me mental health problems (well, probably contributed, anyway).  Application software in the 21st Century really should have garbage collection.  Java I quite like in a way but it's more ridged and verbose than is really necessary for most software of a size relevant to this forum.  Ruby is powerful but feels like a pile of stuff and a lot of the power seems to be aimed at the programmer's foot.

Python, though having a few encumbrances from its rather simplistic roots, seems pretty clean and is now widely enough used that it isn't going to be just a passing fad.  Learning it is probably a good investment for anybody likely to be involved scientific, engineering or server-side-web programming and probably most other areas, too.  The availability of a lot of libraries, many good, is a big advantage.

[KenB]

StB

Incidentally just back from holiday and 3 of Ken's recommendations for Chinese flow sensors have arrived.

Have you had a chance to try them out?  How do they look?


Ken

[StBarnabas]

StB

Incidentally just back from holiday and 3 of Ken's recommendations for Chinese flow sensors have arrived.

Have you had a chance to try them out?  How do they look?


Ken

Ken
not yet. I need to figure out how to marry them to 15mm standard copper pipe. Are there any adapters you know of? Hopefully they will be more "plug and play" than the RS ones which need the LED current carefully tweaking.

Sean

[KLD]

Sean, are they the same sensors that Ken referred to here? Then a 1/2" BSP female to 15 mm compression fitting (BES part # 6655) should do the trick?

Klaus

[StBarnabas]

Klaus
that looks the ticket! Have ordered a few and will let you know how I get on.
Sean