I can not, for the life of me, tell what this is from the website. The video seems like a parody from the HBO's Silicon Valley show. I even clicked through to the app http://app.flowhub.io/ which wants me to log in and then clicked on the "Docs" Button and finally got to this:
Flowhub is a web-based IDE for flow-based programming. It is built on NoFlo.js for both client and server. It can connect to any language or environment that can talk the FBP Network Protocol.
Which is kind of nonsense unless I'm already clear what "flow-based programming" was. Which I wasn't until I google'd that. Maybe its me. It seems if you're going to promote a new tool with for a obscureish programming paradigm you might want to explain the paradigm as well.
You're not the only one who had no clue. I still don't. I watched the video. I made this comment. And I'm not wasting another second of my time on flowhub..
An implementation of TodoMVC is really a good idea! There is already a (re)implementation of Jekyll using NoFlo that you might find interesting: http://bergie.iki.fi/blog/noflo-jekyll/
Granted that dataflow programming (DFP) is probably a bad idea for general-purpose software development, I still think there are some great use cases for it.
For example, I think it could be a great fit for business intelligence and data analysis, since DFP really is a pretty natural way to think about data transformation and aggregation, and most people that would not be able to understand a map/reduce function or a complex Python notebook would probably be able to understand a flowchart-like diagram showing the transformations applied to the incoming data.
Business process automation could be another great use case for this technology, since it's also natural to think of a business process as a flowchart.
The advantage of these specialized use cases is that you don't have to reinvent programming from scratch to make them work.
Enterprise ETL, process automation and service integration is a massive market, it'd be cool to see them buddy up with someone like Mulesoft. I tend to think these kinds of services aren't really suited to general consumers but there's definitely a place for it in the enterprise space.
Sure, there's space for graphical flow-based programming to wire enterprise components together—but there also a slew of established enterprise products occupying that space.
Woah, I have no relationship whatsoever with this guys, but I am quite disappointed with the kind of responses in this thread.
In my opinion there should be no destructive critics in any "Show HN" thread.
If you find some mistakes or whatever, you could make a constructive critic, ie: Hey guys, IMHO you should add more info in the landing page, because I am not understanding what your project does.
On the other hand, as someone posted in this thread you could tell them: "You're not the only one who had no clue. [...] And I'm not wasting another second of my time on flowhub.."
Is that the kind of answer that a fellow hacker deserves?? I don't think so.
Remember that people do not post projects at HN only to reach a big audience, but also to get a great feedback from smart people. So before giving impolite answers, we should remember that we are all in the same boat, and the next "Show HN" thread could be your project. Think about the kind of answers you would like to get.
PS:I know that most of the people at HN is great people, but I think this needed to be said today.
If you Show HN, and someone takes enough time to give it, that's valuable to you no matter the tone.
"Didn't get it, not wasting any more time" is an incredibly valuable comment to receive because it helps you understand the most common occurrence for your home page, someone landing and immediately moving on. You'd normally never get feedback from that traffic and not know why they bounced. Now the builders know what to fix to increase engagement.
IRIS Explorer was awesome, but hardware was not really up to the job in those days. There was also a rich market in GIS and other applications that did what you needed so there was no way you were going to roll your own with IRIS Explorer.
However, the scope and ambition to do stuff with it was incredible. It was far beyond the flat (as in non-3D) internet we have ended up with.
IRIS Explorer was 'flow programming'. You did not have to use it to overlay a vegetation map over a terrain map to render it in 3D and step through the seasons, with widgets to control al aspects of the show, you could use it to render something lame if you wanted to.
Despite the awesome-ness of IRIS Explorer it failed in the marketplace. This was not because of hardware or that there weren't milliions of code modules to casually wire up. For me, trying to do things with it, the problem was because of the 'black boxes'. Sure they had pretty controls on them and you could pump your data in and out of them with a few mouse clicks. However, at some stage you had to go inside those black boxes and work out what was going on. A box could have '2 + 2 = 4' inside it but there was all kinds of toolkit code to take the inputs and set the outputs. A couple of lines of code would do, but those black boxes had a lot of cruft with them. All very off-putting. So, if 'flow programming' with flowhub is to get anywhere, there needs to be an easy gradient between the 'as per the video' way it works and the reality, i.e. inside the black boxes.
IRIS Explorer seems really interesting! I agree with you about that "loose of control feeling". I think it is a common effect that happens every time we go to a higher level or alternative paradigm. We have to create ways to easy go trough those levels and back again.
With Flowhub you can always go inside the 'black boxes' creating your own components as reported in the "Component Editor" section of http://bergie.iki.fi/blog/flowhub-beta. It is also possible to create and remix subgraphs that is another way to abstract a complex group of components.
I believe the British 'NAG Group' bought IRIS Explorer and that it is still being used in academia. You should approach them, explain what you are doing and try to get a demo.
Reason being that wheels get reinvented, and, in IRIS Explorer, there were lots of ideas that were far ahead of its time. Some of those ideas have been lost and some bright spark just needs to pick them up again...
I agree wholeheartedly with being able to go inside the black-boxes, and the transition be smooth.
In NoFlo, one of the the main runtimes for Flowhub, the components are JavaScript/CoffeScript|. They can be viewed and modified from within the UI. The API is pretty simple, and was improved further with NoFlo 0.5.
I'm experimenting with some ideas from the functional and automata/finite-state-machine world on how we can improve things further.
I also think it is important that people can freely chose how much and on what level you use text/code OOP and which level you use nodes/graphs FBP. Embed NoFlo as a DSL for a particular problem area, use as general purpose language, or just as a coordination layer between your high-level modules.
| In MicroFlo, another runtime for microcontrollers, they are C++
I think well-structured and designed landing pages are too infrequent, so I like the OP on that alone...and being an optimist, I'm always a little excited when I hear of a new paradigm, or an iteration of it...because that means that there is a (slightly) new way of thinking about and doing things...maybe even better than my own. Either way, I like that the opportunity exists (even if I never get around to thoroughly investigate it).
And I know a lot of this is promo-copy, but this kind of stuck out to me:
> Focus on the right areas without fear of constant conflicts, code reviews and lengthy onboarding processes.
Again, I'm not experienced in the flow-paradigm...but I just have to guess that the need for code reviews is not because traditional code is not visual enough. In fact, I'd argue quite the opposite... a visual diagram offers some great big-picture benefits, but at the loss of granularity...if anything, I'd think a flow-based paradigm would require as much code review as before, but perhaps with a different mindset.
I also think the background image for "Drag, Drop, Connect, Build" is...not inspiring to me. I mean, it's useful in some respects for overview of dependencies, maybe...but dragging-dropping and interpretation of visual symbols is not always more efficient than pure textual concepts. In fact, I'd argue that it can be even less efficient. How easy is it to examine the underlying abstractions behind connecting one pipe from one icon to another?
If the website's performance is a promise of the performance of the application, I don't really feel tempted to try it.
I'm using the latest Firefox Aurora release and I am getting very sluggish scrolling and slow loading on the main page. I wish this trend of gigantic slide-able pictures and animated transitions would stop. Yes, it feels like I'm browsing through a tablet/phone app, but I'm on a desktop, I'd love to have a desktop experience.
Sorry for the rant, the application itself seems nice though.
EDIT: also please don't force people to register and login to try the app, if I can't just get on the page and try it without giving away some private data, I don't really feel tempted to try it at all.
Nah, that is related to GSS stylesheets on the site making use of the matrix3d part of CSS. Unfortunately lots of (expecially Linux) 3D drivers are not exactly up to the task.
Edit: as for required login, that isn't needed to play with just the examples.
I can play 3D videogames with lots of advanced shaders and even develop my own, perform very demanding computations on my 8-core machine and watch HD movies and stream from the web, yet I can't easily view a simple home page from the internet?
Yeah, it is a shame. I'm getting some checkerboarding there too as my Chromebook Pixel can't always keep up with rendering the surfaces when scrolling quickly.
Funnily enough my (a lot slower) Nexus 7 scrolls much better. That is what points me towards 3D drivers.
But anyway, shall we talk about Flowhub itself? ;-)
> also please don't force people to register and login to try the app, if I can't just get on the page and try it without giving away some private data, I don't really feel tempted to try it at all.
Flow-based programming is already common in enterprise finance. It's a large part of my job. Enterprise fbp solutions are extremely expensive and sometimes secretive (ex. Ab Initio), so it's interesting to see this become more mainstream.
I need a better introduction to understand how flow programming works, and how I might use it. Instead I see a (seemingly) complex graph which I understand represents a program, but it's not clear how this really works in relation to text programming. I could spend hours and hours digging to figure it out, but that could be a waste of time, so I won't do it.
I get text based programming. Find a way to quickly connect that experience with flow programming, and conversion rate on this site will improve.
I'm a backer, but I'm outraged with the paid plans. I'm not sure how people design their pricing models, but often they are completely disconnected from reality!
Well, there is also the free plan for open source and public projects. And the UI is open source for those who really don't want to pay for the service.
Flowhub is a web-based IDE for flow-based programming. It is built on NoFlo.js for both client and server. It can connect to any language or environment that can talk the FBP Network Protocol.
Which is kind of nonsense unless I'm already clear what "flow-based programming" was. Which I wasn't until I google'd that. Maybe its me. It seems if you're going to promote a new tool with for a obscureish programming paradigm you might want to explain the paradigm as well.