Low code pains and problems
Hi All
I'm building something in this area and am interested to know what problems and drawbacks people have with low-code / no-code, and what limitations they run into
Trending on Indie Hackers
Hi All
I'm building something in this area and am interested to know what problems and drawbacks people have with low-code / no-code, and what limitations they run into
we've been using https://Wappler.io (low code, visual web app builder without any platform lock-in) for web app development full time for about 2 years now. we're a small team of full stack developers.
the only real limitation we've come around is inability to build a multi-tenant web app with Wappler.
all of other custom requirements we were able to build by leveraging third party PHP/ASP libraries and/or APIs - we've been able to make it work well with Wappler.
What I mean is when you say Wappler is without any platform lock-in, many low/no-code platforms will not make it super easy to integrate their platform as part of a larger project. I was asking if that was an important feature of it for you?
for our clients, most important was to be able to host the app themselves without any 3rd party interference. plus integration of literally any PHP/ASP/NodeJS library and any RESTful API is possible with Wappler - because it is low code - we can access the full code directly at any time.
this is very important when building enterprise solutions which often require heaps of custom integrations.
still using low code is better because a shit ton of boilerplate stuff is available to tinker with 'visually' - hence it is really fast to develop and a breeze to maintain with the 'visual tools' Wappler provides.
Yeah that looks really interesting, it's a great approach. So Wappler is essentially a code-generator desktop application that builds out applications for you in PHP/ASP.Net/NodeJS? Don't you then lose some or all of the ability to use the low-code interface if you modify the generated code yourself? That commonly is the issue with that kind of scenario.
Wappler does not generate PHP/ASP code. it has a JS framework called App Connect which handles all the server side stuff including DB connectivity and API calls. what it generates in code is pure HTML+CSS+JS.
whatever changes we make on code level, most of it is within the Wappler's framework (with some learning curve, its easy to pick up - am an accountant by education but am able to do basic coding on Wappler now after 2yrs working on it.) - so App Connect can parse it just fine.
the 3rd party integrations are done as separate code blocks on the Wappler pages (.html/.php/.asp) or completely on the server side - so it does not interfere with the Wappler setup in our experience.
so in theory, if you decide not to pay for Wappler license to use their UI, you can still read, understand and tweak the vanilla JS code to get things to work - not advisable obviously, paying for Wappler's license is worth being able to do all the tinkering using it's visual tools. but you could surely stop paying the monthly fee and start again once you want to make some changes or anything - your app hosted on the cloud does not stop working if you stop paying Wappler.
as a no-code tool, Wappler is not used to the best of its abilities - but it has that kind of feature set to be able to work as a no-code tool.
as a low-code tool, the power is just immense - coders can catalyse the dev process with Wappler to unlock efficiency gains and custom business logics to enable the 'citizen developers' to focus on being able to build the workflows purely through visual tools.
it's a bit to grasp! but if you have something in mind, would be happy to show you around how Wappler works on large scale enterprise projects with custom business logic and small scale basic CRUD apps with minimal custom business logic - we've built both.
Yeah Wappler looks good! So you find it's great in terms of being easy to use alongside custom code and other products and that stops you being limited?
Didn't follow your question - can you clarify please?
Hi James,
I'm wondering what you ended up building? What are in your opinion the most common pain points besides what was mentioned in this thread?
Hi Alexandre, haven't been on here for a while so just saw your reply! I'm building this: https://restspace.io. It's a kind of low code which works by supporting intensive code reuse, thus allowing you to build a back end mostly out of prebuilt mini-services. But still allows you to build your own services really easily. I'm adding a function to cross the barrier between front end and back end in NPM modules, so that an NPM module can contain a back end specification and services, and the platform can automatically build what it needs: a kind of full-stack package. So e.g. you could have a user registration package in React containing a spec that could automatically build the back end it needed to fully operate.
I think another issue not mentioned here is where a no-code platform to become sufficiently capable to deal with complex problems also becomes hard to learn. But the effort investing in learning it becomes locked-in to just that platform whereas if you'd learnt e.g. javascript from the start, you can then use it in a wide variety of contexts.
Also mentioned here is self-hosting is obviously not generally possible so you're tied to your provider's SLAs and speed of processing. And commonly there comes a point during a business scaling at which it will need to move it's application off a no/low code platform which could be quite painful (but could also be advantageous if it's handled right in terms of your version 1 of custom code has a lot of knowledge of what the application needs already learnt).
hello, great questions. often we encountered some limitations with "looping workflow/process", example : we need to go thru some results, integromat cannot do it in our scenario or the data manipulation is pretty hard. it's often one linear workflow.
We are working on this problem at Phiona due to the limitations we had with existing automation tools that only seemed to work with single events.
If you don't mind me asking, what type of data transformations do you need to perform on a regular basis?
That sounds interesting, what's your use case where you need to handle multiple events ?
A lot of the use cases we've run into are subsets of existing ETL use cases, like moving data from Google Sheets into a database (or vice versa) and making sure the data is in the right format when you have anywhere from dozens to millions of rows.
In particular, things like moving individual employee timesheets to a company-wide document, or updating inventory levels of different stores or restaurants, seem to be the areas of most interest. Too much work to do manually, not enough work to build a fully coded solution.
So you find it hard to get powerful enough data transformation? Specifically with integromat? Interesting to know what they have in that area.
exactly, pipedream also, you can define a workflow but cannot doing loop thru some data, and yes, transformation. from my last test, integromat was fine for some scenarios. but, transformation, you need to use about 4-5 modules and it gets pretty heavy scenario. instead it could be nice to have a kind of "code snippet" where your can do what you want with simple line of code and continue the transform with your output.
OK so the pipeline module in what I've built can split an array, or set of files apart and process each part in parallel, would that solve your use case?