What needs improvement with Mendix?

We are all moving away from a monolithic product model to microservices. We are building an F2DUI application to decouple the front and back end. Mendix provides an integrated approach for both. Ideally, we should have a way to separate the front end and back end to scale them independently. While Mendix does offer different ways to achieve this, I sometimes prefer having the front end separate. I don't want the back end to be tightly coupled with the front end all the time.

The vendor should focus more on the opinion of the users and make improvements to the product accordingly.

The code refactoring tools could be better, especially for applications running for years. It's not bad, but it could be smoother. Also, writing new widgets can be trickier than it should be for some people, but not if you're familiar with Mendix. Pricing used to be complex, but Mendix has improved that quite a bit. So, the pricing policy in general. It's not exactly straightforward. The reporting feature also wasn't great, but Mendix has fixed that in version 10. I still need to test it fully, though. Version 9's reporting could be better. In future releases, I would like to see more features around report writing. Another thing is managing and breaking down monolithic apps into modules, with version control for individual modules, which could be more robust. Sharing individual modules wasn't as strong as it should be. However, the biggest improvement still would be better refactoring tools, both for code and models.

One area for improvement is its integration capabilities. Creating a pluggable widget or integrating it with other systems is challenging. In terms of features, it would be great to see advancements such as AI services and the integration of third-party services. Additionally, connecting external devices to the application requires multiple steps. Improving this will make it easier for the developers.

One thing I would like to improve is the support system offered by Mendix. It can sometimes take a while to get the help I need when I'm using Mendix.

The product needs more connector integration with Microsoft products. We'd like to be able to write in C Sharp to develop code for Mendix. However, that may be impossible. Still, we'd like to be able to modify the programming language.

Feature-wise and in terms of technical aspects, Mendix is excellent, but its pricing is steep.

In the beginning, it is difficult to learn and work with. I have to say it was difficult in the beginning to understand best practices. Once I got the best practices for the back end part, for manipulating data, it was easier to use. For the front end, it was a bit more difficult. For the front end part, I don't like that it's not fully drag and drop there, and I have to know a bit of CSS, which I don't really know. It would be cool if it could be completely drag and drop, and you could do it like other products, like PowerApps. With PowerApps, you can move it pixel by pixel. It's like PowerPoint. I've experienced a few bugs. Sometimes it was really tiny things - not something big. There are issues with the versioning systems, et cetera. It had errors deploying in the public cloud to publish a solution just for the sake of testing it. I had an issue with a message definition. When you create a message definition and save it, and then close the project and open it again, the message definition did not save. I had to do it again. This has happened a lot.

The cost model could be better. It is expensive. There should be more integration with engineering applications and tighter integration for user authentication, such as single sign-on, etc. They have some of that. It just could be stronger.

I struggle with solutions like Mendix in terms of creating enterprise solutions. When I say enterprise solutions I mean enterprise-grade solutions. Let's say if I create an application on Mendix and I want to roll it out to multiple countries, that kind of thing I can do with Mendix. I'll have to copy it. That's hard to do at an enterprise level which can be quite sizeable. I'm not sure how well it scales. The setup itself can be complex and difficult. The UX capacity is lacking. I'm not sure how well the workflow capabilities will hold up.

Mendix could improve by allowing the customization of different programming languages, such as Python and C++. In a feature release of Mendix, they should add 3D augmentations or other 3D visualizations, such as images. Additionally, more customizations would be a benefit.

Mendix needs to think about itself offering machine learning and artificial intelligence. That's going to be the future. I really like that they're already working on new features. Nevertheless, to really be on top of things, they will need to do more in artificial intelligence and machine learning.

There is always a layer of custom code required. There is a misconception of low-code, or Mendix, or the industry in general. They are perceived as more of a dashboarding tool, and as a visualization platform only, rather than building a complete enterprise solution. That's more of an awareness marketing challenge they have, or the industry has. In general, AI needs to be better. The team and the company is running ahead with this a bit more. AI area is something which companies have started to pick up on, low-code wise, and they should invest in it more. I would like to see their data hub module become a little bit more mature. They need to expand their base as the concept is amazing. We just need to see more use cases and learn more capabilities there, and then definitely they need to fill in the AI piece of it.

I'm still in the evaluation phase. I have not fully vetted the entire product yet. It is expensive.

Students do have some difficulties translating the hard code. They are usually using code from a previous course. My understanding is that, if you are not using the free version, it is very expensive. We'd like more support and more publication of use cases and examples so that students can more easily study the product and better understand how it works and its applications.

The cost of total ownership needs to be better. The licenses are very expensive. If you compare it with, let's say, a kind of BPM, or CRM solutions, the cost is very, very big. While the community is great, they need to work on making their direct technical support services better.

We would like to see is the ability to version manage modules and not just the app. We need finer-grained version management for software repositories. Version management is good but it needs more work. Also, because of the licensing model, Mendix apps are too monolithic. It would be great to have a microservices licensing model that works well for microservices especially designed to work with Kafka, Google Pub/Sub and streaming technologies. We need much better code refactoring tools, like IntelliJ but for Mendix. For example, if I wanted to maintain all projects and refactor Domain model fields in bulk, it would be good to have intelligent renaming across the whole model with regular expression syntax. The modeller is too clickety-click. The most important feature I'd like to see is support for is first-rate JSON Schema support and first-rate GraphQL support. Of course, the Modeller must run on Mac, not Windows. Lastly, the licensing model does not scale well with many users. This is a huge problem as apps that have many users become very expensive and can kill the business case. Mendix is reasonably negotiable but it's a pain to deal with the licensing for each new project.

In terms of what could be improved, of course I'd like it to be highly secure and highly scalable. Security is paramount to us. Scalability-wise, we don't necessarily have a huge volume. From a scalability perspective, at this particular stage, it's not necessarily a top priority.

While it is difficult for me to address which areas of the solution are in need of improvement, as I manage Mendix projects, but do not utilize them, I would point out the transformation of the BPM and the process with other applications. This is because the applications are conceptually what we need, as they concern the RPA Automation Anywhere and UiPath. Mendix has the potential to redesign its process engine for compatibility with a BPMN tool without any further action being required, enabling seamless integration between the business processes design applications. As mentioned, there needs to be an increase in the number of the UI components so that other platforms will not be used in place of the UI interface of Mendix, such as Angular. Improvements should be made to the functionality to increase the number of UI components in the actual software. As such, we occasionally go outside the system to develop certain pages in React, Angular or Vue JS, which we then import to Mendix. This we do because of the limitation in the UI design, but it is not on par with what's happening, for example, with OutSystems or other tools.

There's a new update coming soon, and that will be full of great items. It's not so much that there's room for improvement on the product. They're creating some custom or some out-of-the-box modules that are going to be a part of it. In particular, they've got a workflow module that we could replicate-build ourselves, so to speak. It's probably a module that would take a couple of months, and then you can tweak it. To have that out-of-the-box potential for certain aspects is going to be really good. Having all that workflow prebuilt will be amazing. There's no direct tech support. However, it's not the type of product you really would get tech support on.

We specifically want version control of whatever we deploy in the production. However, with Mendix, unfortunately, storing things in binary is quite a lot. If you put it into Github and stuff like that, we'd like it to be a text type of source code rather than binary. There are not enough developers who are using Mendix. The knowledge base available online and in the market is not as rich as other competitors. In terms of workflow automation, local automation, they should focus on integrating AI modules. Many companies are trying to build their own models, however, if you compare with how much learning Google has done on their Google models, it makes sense to go for integrations rather than starting from scratch with teaching your AI. That's roughly how we see the domain of RPA and local automation platforms. The integration or making use of some artificial intelligence and models which are out there could be a lot better. There should be some kind of marketplace. The user interface styling is a bit tricky. It's got a restrictive and highly sophisticated styling which could be better.

You need experienced programmers and developers to understand this solution. We had a very experienced developer use the solution and they had difficulties, the training for developers could be better.

Overall, integration with the enterprise ecosystem needs improvement. I would like to see the inclusion of APIs that can help with the interoperability.

I would like it to help us be more productive. Also, while the documentation is good, the development box could be better.

Mendix is right now focusing on developer features while we would like to have more features for the information analyst. Mendix is not going this route because they clearly see the developer as their target group, while we like to see it as more of a tool for information analysts. What is lacking is the support of higher level modeling features, like the modeling you do is relatively low level, yet it is still close to programming. We would like to see a more business-oriented modeling environment, like BPMN. However, that is not the direction Mendix is going. We are developing our own intellectual property in this direction. Thus, we are building it ourselves.

The document templates definitely need some love, they have been around for a while but no actual improvement have been done since they were released. At least not in terms of additional options like the web counterparts of the elements you have available have. Also, Mendix is even easy for non-developers to start developing apps with, however when the applications grows and gets more complex these type of users are not fully guided in making it as secure and performance as it should. It would be great if the modeler could also start pointing in the right direction when it comes to that. Although Mendix did introduce the Mendix Application Quality Monitor which could help with this issue.