Hacker News Reader

Comment by sandGorgon | Hacker News Reader

please please use an ORM. if yoy dont build database abstraction today... you will never be able to build it once the plugins take off. Just the fact that you can run postgresql will be a killer feature.

i think the golden usecase would be running thirtybees on heroku (postgresql, composer, php 7)

EDIT: there was a dead comment about how postgresql is not good. what im asking is different - im requesting for an ORM like Doctrine or Propel. This will let the end user choose the database that they want to use. But you have to take this stand pretty early in a project's life cycle or you will not be able to do it later. Exactly the issue with WordPress.

http://www.doctrine-project.org

http://propelorm.org

View on HN

I get it, but honestly it is not in the immediate roadmap. Its really not something that people in the scale we are in are worried with.

an ORM will help you write more maintainable code in general. Every framework right from rails to symfony uses an ORM for the abstraction it brings.

Some of the security issues in wordpress - unescaped queries, SQL injection and everything - which plagued it for decades could have been avoided by using a well tested ORM.

I would urge you to make this one of the highest priorities in your code cleanup. It is not going to take you a lot of effort, but the long term advantages are too many to enumerate here.

For example, for your own development sake - I dont know how you plan to manage schema changes. If you use Propel, you will use "database migrations" - something that every framework from Rails to Django, etc mandates as best practice.

I understand it, but look at it from this point of view. PrestaShop 1.7 requires everyone to purchase new themes and modules. The code base is extremely messed up.

We are taking the PrestaShop 1.6 codebase that has 10k modules and 2k themes as a fork. If we go changing the database handling off the bat we are going to lose the compatibility that might make us successful. We aren't a project starting from scratch with unencumbered code. We need to maintain compatibility to be successful in the beginning. We are springboarding basically. You are wanting us to take the springboard away and just jump.

hmm.. i wont say I'm not disappointed. i thought you guys were going to rewrite a lot of the core because everything is broken anyway.

since 1.7 is already breaking backward compatibility.. maybe it was an opportunity to do it right.

A simplified version of the roadmap is this. We are going to fix bugs and stabilize the platform where people can sell without having daily bug fixes.

After that we are going to pool our knowledge, resources, and community to figure out the path forward. I will be totally honest with you, I am half developer half CSR person. I hear what my clients want and I try to make it happen. I am not the only person in on this though. I am one of many and very flexible. We are going to do a rewrite after we have stabilized things and we are going to do it right. We are going to look at the options out there and figure out where to go.

At this point it looks like I am leading things. I am the reluctant leader. I am the guy that realizes that my ideas might not be the best ideas and I want to hear yours. I just feel in the beginning it is about stabilizing what is there. I think if we spent 6 months rewriting everything we would lose an important window.

Tonight I am actually setting up the feature voting for our site. This is how I think things with a community need to be handled. There are going to be nay-sayers that stifle things because they are new and unfamiliar. This is always going to happen. But if we are transparent and add features by simple voting then no one can really argue with the logic. I have my ideas, but I am not the arbitrator of what will be included. I really want this to be a community project. I want the merchants and the developers to speak what they want and I want to make something that reflects the ideas of both.

You should worry about it, and fast, as every professional and actually usable and maintainable setup uses one, and your refusal to use one, thus relying upon a patchwork of unknown stuff from unknown sources, is one of your biggest exploitable (as in I do it to clients while testing their network every day,) vulnerabilities.

So far, I've moved a lot of people, including my own online auction company, off of your platform because getting AJAX code to work reliably with it is near impossible due to lack of object-relational mapping.

I could go on and on, but quite frankly, you need to start from scratch. Right now you're just trying to slap a lot of icing on a poorly-made cake.

Reply on news.ycombinator.com