Skip to main content

I think we can all agree that FME is pretty unique, special and powerful in it’s no code approach to data movement, integration and automation.

I’m sure many of us have heard of, or even used similar (FOSS) platforms to FME, such as Node-RED (https://nodered.org/) or IFTTT (https://ifttt.com/)
 

‘The algorithm’ recently pushed n8n towards me - https://n8n.io/

 

These platforms are more focused towards the API/Event integration side of things, rather than the entire data stack that FME excels at. Of course, with the right workspace, FME can do everything these tools can (and more)!

I don’t particularly see these tools as being much of a threat to FMEs core data manipulation as they’re primarily targeted at stringing web events together.

What I do think however, is it’s opening the doors to developers/teams/companies who have previously only ever thought of lines of code being able to solve solutions. It’s raising the awareness of these no code solutions and how (GUI based) tools like FME can enhance and improve workflows across organisations.

 

I don’t really have a question, but what thoughts do others have on this?

Thanks for the interesting video, Hamish! I hadn’t heard about n8n.

I agree about event integration and automation getting more and more important, not only between applications/platforms, but also through MCP servers that allows AI to interact with external systems such as databases and service providers. 

As a site note, I remember using my first graphical no-code development platform in the early 90’s when using LabVIEW for processing real-time data from the serial port. As is sometimes the case, this was great for getting a quick prototype up and running, but in the end we had to develop our own software from scratch because of limited functionality and not enough fine-grained control in LabVIEW. The licensing cost was also a major issue when we wanted to distribute the solution to clients on a larger scale.

I’m not comparing FME to LabVIEW, they are very different beasts, but there certainly are some interesting overlaps in both philosophy, practical use and (perceived) limitations, perhaps due to the inherent nature of no-code development platforms that needs a proprietary back-end for execution.

I firmly believe that both code and no-code development have their respective, strong use cases and that they will continue to live just fine side by side: Sometimes you need to get up and running fast, and sometimes you need to eek out the last bit of performance or functionality, development time be damned.

Where I’m really interested to see what the future brings, is at the intersection of these two, where you can get the best of both worlds.

 


Fireship is a great YouTube channel for tech news and for popularizing technical knowledge. I would recommend following him !


I’ve used IFTTT a little bit in the past and it just so happens that we were visiting a client yesterday who told us that “another FME partner” was setting up a system for them where FME Flow’s scheduling and automation capabilities were used to run python scripts 😳

I’m sure we’re all convinced of the benefits of FME due to its graphical interface and I hope more of the hardcore Python programmers start seeing this way too.

 


This tool was actually prompted me to have a go at this:
 

I saw a nice video of someone playing with n8n to build an agentic workflow and wanted to have a go with FME - I must say it was a bit painful and I found myself envious of the n8n workflow/ui. It’s been on my list of things to play with for a while now.

I would love for Automations in FME Flow to be expanded to this kind of scale. At the moment Automations are more of a way to automate the running of workspaces. 

I would love to see a world where Safe have the FME Flow core be free / open source and keep the engine as it is - I don’t seem them doing that though. 

I also really like the Blender/Unreal Engine approach (and likely many other) where the UI components are simply just blocks of python or C++ where you can build your own. In a sense the PythonCaller allows for this, however, I can’t just take a reader and tweak it to add support for a missing functionality. That would be amazing.