Using an ORM?

[Aleksandergreg]

In regards of this from Anders:

Following the DevOps principle of showing your work the discussions regarding ORM and migration tools should be made public on Github Discussions.

[Aleksandergreg]

I don't think we should use an ORM, since the queries we are making are quite simple, and only in two different tables.

The only reason I could see for us using an ORM would be to know how it works in Ruby.

[theaskou]

I agree with Aleksander. After reading about the benefits of ORM, I can see that it would give us some advantages, but compared to the scope of our project, and how much time we would spend on implementing an ORM, I am not sure it will be worth it. And since it is also not a requirement for the exam, I think we can spend our time on something else. I wouldn't mind if you would like to try to implement an ORM though.

I guess it depends on how much we expect to expand our operations with the database. After we discussed the missing data in our KPI, I can see that we might need to store some more relational data, like logging interactions, last sign in, or similar. What do you think?

Another type of benefit of implementing an ORM would be to even out the contributions to our GitHub repo - Aleksander has made far more than anyone else on the project so far.

Here's what I read about the pros and cons of using an ORM:

Pros

• Makes query writing easy and fast
• Reducing boilerplate can make developers more effective
• Makes it easier to switch databases
• Better readability
• Easier to maintain
• Many ORMs provide caching mechanisms that store frequently accessed data in memory, reducing the need for repeated database queries.
• Lazy/Eager loading
• Inherently protects against SQL injection

Cons

• Lower app speed performance due to the abstraction layer and extra operations
• Learning curve: It might take us more time to learn, than it would benefit the project
• Tends to be slower, requires some load management to avoid this

Does any one have an idea of which ORM would be good in our case?

[theaskou]

... I just read this from Anders' slides. I think that settles it 😅

"In this project, given its scope, fixed deadline and data complexity an ORM does not make sense. There isn't going to be more functionality on the serve from this point on.

I want you to be aware that it is not always a good choice.

An ORM is either equal to or less efficient than writing raw SQL. An ORM is not "more secure".

When creating a small project or MVP, developers shouldn't waste their time on setting up an ORM.

I will accept during the exam that the reason you set up an ORM was to learn how to use it for the language you are using."