Return to “General”

Post

Visual Node Programming (Dev Log Oct 23)

#1
What I really want to do is make a tool where code is built visually and structurally. Something like node-based code, where code becomes this simple, beautiful web of structure. Function and data would be the same thing - just nodes, and structure would be created visually rather than syntactically.
I think every long-time programmer had this (similar) idea at least once.
It's imprinted into the programmer's head / DNA and part of the mass group conscious if you will :ghost: :D .
There must be something better! This can't be everything in (a programmers) life!
You could call it the holy grail of IT - if you really can make this one happen as part of a general programming language, you would never need to buy yourself drinks again when Nerds are in the same bar :mrgreen: (as if that would happen :D ).

There are some high-level, high earning programmers who have quit their job since its just sucking to try to 'talk to machines' in a way which is un-natural and so prone to errors.

Many have tried to build this node-language though - I know at least 4 projects which are all done for individual cases - but they all fail: they put up the 'bad strings and texts' in those nodes again. :evil:

I guess you need to be able to actually imagine this kind of concept - and this is which is really a problem since it will be VERY hard to come up with this stuff for general purposes.
If you can do it, Josh... do it. I just think that you will need 3x the time as for LT to do it right, but you could do it!
You know you would shape the global landscape of IT. Many will (have to) change jobs. :?
Post

Re: Visual Node Programming (Dev Log Oct 23)

#2
Isn't it just an extra layer? Perhaps even a bit of an API for programming?

I think it's indeed the next logical step in the evolution of programming though.
Perhaps not really the visual node thing specifically, but making programming more like a real language. Perhaps even a bit like on Star Trek. Tell the computer what you need and it builds a program for you.
Beware of he who would deny you access to information, for in his heart he dreams himself your master.
Post

Re: Visual Node Programming (Dev Log Oct 23)

#3
kone wrote:Many have tried to build this node-language though - I know at least 4 projects which are all done for individual cases - but they all fail: they put up the 'bad strings and texts' in those nodes again. :evil:
Very interesting, glad to know I'm not the only one that thinks it's annoying how we do it!! I figured other people get frustrated with it at some point too :D

Would really love if you could link some of these projects? I've search around a bit but can't really find anything relevant. Mainly because I'm not sure what terms to be searching for, or what this kind of thing would be called if it exists. Would love to see what others are working towards though :)
kone wrote:If you can do it, Josh... do it. I just think that you will need 3x the time as for LT to do it right, but you could do it!
You know you would shape the global landscape of IT. Many will (have to) change jobs. :?
I honestly think I have to, at least for myself. I won't claim that whatever I do will be good for anyone other than myself (as I know that I'm personally heavily biased towards visual / spatial / structural thinking). But I can't keep doing this text-based thing for my whole life, I'm sure of that :shock: Maybe someone else will do it though, and I'd be equally content with that. I just want my code to look as simple as it deserves to look :(
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford
Post

Re: Visual Node Programming (Dev Log Oct 23)

#6
JoshParnell wrote: Very interesting, glad to know I'm not the only one that thinks it's annoying how we do it!! I figured other people get frustrated with it at some point too :D
Very frustrated!!

Code: Select all

Since the first year of studying programming at university I have known in my heart that computer programming is not meant for me, but I was afraid to do anything about it and here I am now 12 years later programming with no passion. I am a career programmer and an average one at best.

I come to work every day with no passion I just do it to pay the bills. I have done some good projects but I am not at all into it.
read it in full: here

I have seen other articles which tend to describe that the programming itself could be better.
I once read a very nice post about the 'modern guys' who find that one new programming language promising everything you need. They drive the maturity, they write books, and once it's a mainstream language they move on. So that's not a solution. :think: Also it's no solution to say: Ruby is better than C# or something like this.
Although the code by the 'new' languages might look like its getting shorter and shorter (since the built-in tools get better and better I think).

Btw. you might want to grab a Code keyboard from this guy. :D
JoshParnell wrote: Would really love if you could link some of these projects? I've search around a bit but can't really find anything relevant. Mainly because I'm not sure what terms to be searching for, or what this kind of thing would be called if it exists. Would love to see what others are working towards though :)
  1. See this one for Visual Rules.
    It's based on eclipse (I think) and is a modeler for business models.
    In this screenshot you can see a model for a car rental company.
    The idea is that with this visual concept also a 'normal' business user can 'code'.
    The model gets compiled into .jar files and can be used as models in a platform running on a Apache Tomcat or JBOSS or similar.
    When really 'running' it though, you are limited to typing in numbers into the input-fields and out comes the output (which the model is calculating). :monkey:
    Of course you can automate this and put it via automated interfaces in the systems at hand so that you can actually work with it without problems and without too much manual hazzle.
    They lack two things though:
    1. no real time. You need to compile the model, then submit it to the platform. Of course you can link it, but well it's just not the same...
    2. No really good reporting. You simply cannot see how your changes will affect the future given the historic data. You need to actually run those batches producing reports and then use an external reporting tool to see whats changing on a global scale.
    Have a look at the full screen link. Image
  2. Another one is a flow language I cannot disclose since there are no pictures of it in the web. If I extract from the systems I have access to it would be a violation of several contracts.
    But let me describe a little: It's a calculation tree. It serves as a drill-down feature too.
    Let's say you see a financial number calculated. You press the right mouse button on ANY number on the screen since you need to know how the hell this one was being calculated. The drill-down calculation tree pops open and you can click deeper in the tree. For each element you see a 'black box' with the input variables/arguments, the output variables and (if you are lucky) a description of the function.
    In seldom cases the black-box function is wrong. Often it's the user who didn't use the right arguments.
  3. Next up is Lego (Mindstorm) NXT.
    It's much more powerfull as the normal person might expect.
    Sure it's powered by a small chip only and they only recently added support for WIFI - but you are now able to do really astonishing things with Lego Robots (or Automaton if you will :ugeek: )
    You can do nearly anything with it from robot fights (push the other robot from the table) to automated quests such as: find the light-emitting object, bring it to a location etc. etc. - you surely could do a good vacuum cleaner bot when lego would finally release the suction motor. :P
    The software is really neat and easy to learn.
    I found no better picture so I leave you with a video where a small child explains how to program a robot following a line.
    If you don't like being educated by a kid: watch this guy.
JoshParnell wrote: But I can't keep doing this text-based thing for my whole life, I'm sure of that :shock:
You are definitely not alone!!
JoshParnell wrote: Maybe someone else will do it though, and I'd be equally content with that. I just want my code to look as simple as it deserves to look :(
Or no code at all... :ghost: :ghost: :ghost:
Post

Re: Visual Node Programming (Dev Log Oct 23)

#12
Just remembered something: the MIT app inventor. Its kinda graphical development tool for android apps, its main purpose us education and its not very powerful but i mamaged with it to build a bluetooth remote app for a freinds robot in just a couple of hours (mainly spent on iterface building). My friend was mad at me afterwards because he had done it the "proper" way in about 10 weels from zero java/android knowledge, just as me ^^

http://beta.appinventor.mit.edu
Post

Re: Visual Node Programming (Dev Log Oct 23)

#15
JoshParnell wrote:Would really love if you could link some of these projects? I've search around a bit but can't really find anything relevant. Mainly because I'm not sure what terms to be searching for, or what this kind of thing would be called if it exists. Would love to see what others are working towards though :)
A cautionary tale:

A decade or more ago, I worked at Microsoft in its heyday. The company was full of coding rock stars, and it seemed like the stock would never stop going up. In those days, if you notched some major successes, the company would cut you a lot of slack on your choice of project. One such success was Charles Simonyi. After famously making Office a success (and infamously inventing Hungarian Notation), he decided that coding needed a reboot. He formed a team and set to work.

In his vision, code formed a semantic tree, but you never had to deal with that directly. Instead, you would have different UI layers for different purposes. One might look like a text editor. Another would look like an electronic wiring diagram. But they would all be backed by the canonical tree. A complex series of transformation stages would convert the tree into binary code, or (in theory) any other final format.

It demoed extremely well, but as the years passed, each milestone would produce another variant of "hello world", and the project never found a team or a customer to call its own. Eventually the company cut the project loose. Mr. Simonyi still believed in it - as far as I know, he does to this day - so he took the project, left Microsoft, and formed Intentional Software.

Every couple years I check back in with that company site, hoping to see the transformative breakthrough achieved. So far its just been an infrequent stream of stale press releases and unimpressive case studies. But hope springs eternal.

Online Now

Users browsing this forum: No registered users and 1 guest

cron