Has Magento 2 Reduced Integration and Upgrade Effort? The developer beta period (officially commencing December 18, 2014) is a great time to explore Magento 2 and provide your perspective and feedback. What is better? Where can Magento go further? Are there rough patches where small changes could have a big positive impact?
In a previous post I asked is it useful to categorize modules? A part of this article was asking if a “site developer” is a useful role? The feedback I got was more like “this is an interesting idea, but its not what really happens when developing a site in 1.X”. This does not mean its not useful to categorize modules – it can make Magento easier to pick up and learn. It also does not mean you cannot build a site just by adding modules – this can be useful to prototype a store or explore a new extension. But a core value proposition for is Magento is its ability to be a Chameleon – its ability to be customized to develop your site with your branding and your unique look and feel.
So what should you explore in Magento 2 during the developer beta period regarding reducing integration and upgrade effort? Here are some suggested areas to dig into. (The following assumes you know Magento 1 as a basis of comparison.)
The following areas have been changed in the presentation layer:
- A lot of business logic has been pulled out of PHTML files and blocks and moved behind service contracts. Benefits should include there is less code in PHTML files if you want to customize them.
- The introduction of service contracts with easy exposure as REST or SOAP services should make AJAX calls easier with consistent behavior to PHTML and blocks.
- There are now more, smaller PHTML files. This should make it easier to replace individual PHTML files where changes are required.
- CSS has been extracted from PHTML files. This should make it easier to replace CSS without having to replace corresponding the PHTML files.
The following are some of the changes at the domain or business logic layer:
- Service Contracts have been introduced to provide slow changing APIs across releases. Other modules that depend on the service contract are less likely to require changes during upgrades.
- Plugins support has been added to avoid class rewrites. This should reduce the number of extension conflicts as plugins can nest. (In Magento 1 it was not uncommon to rewrite a class – if two modules rewrote the same class only one rewrite would win, requiring the system integrator to manually merge the changes from the two modules.)
- Events are still there as before. I am particularly interested to see if plugins are a good replacement for events, or whether events are the normal extension approach with plugins being available when events do not currently exist.
- Modules have been split up somewhat to make it easier to remove unwanted logic and add new extension logic easier. For example, each shipping method is now a separate module allowing easier control over which shipping methods can be offered.
In many ways this post does not say anything much new. All of the above features have been mentioned before. However my intent behind this post is to get people ready to start thinking about whether the changes as made have gone far enough to really solve the targeted issues. Even better, are there any low hanging fruit where a little more effort would go a long way to solving a remaining problem?
It is not mandatory to fix all problems in Magento 2.0. We can continue to iterate and improve after the release. But this less desirable due to the impact on extensions and sites. So it is desirable to get the base platform as right as possible now. This is what the developer beta period is intended for – for extension developers to explore the changes made and provide feedback.
So when should extension developers port their extensions? I have answered this before, but I think its worth repeating. My default generic advice would be to explore during dev beta (Q1 2015) then plan the serious porting after that. If you want to avoid possible rework then port your extension as late as possible (so porting completes as close to the GA release as you can). If you want to have more likelihood of changing Magento 2 to suit your extension, do it as soon as possible. Some extension developers may also decide to port their extensions earlier to be ready and start training up their own staff on 2.0 to be ahead of the curve. But my default advice is to explore Magento 2 during developer beta (so you can raise issues or suggest changes), and then do the serious port after developer beta to minimize rework required.
Thanks a lot for your series of Magento articles on magento2, its really helpful for beginners like myself.
BTW I tried to install magento2 on windows platform running PHP 5.5.19 and MySQL 5.6.X, using composer, but i received the following errors (http://pastebin.com/WSWrKPif). I would like to confirm whether Magento2 is compatible with PHP 5.5.19 and MySQL 5.6.X. Kindly let me know if i miss some prerequisites for magento2.
PS: Please ignore (remove) my comment if you find its irrelevant here.
Yes, PHP 5.5 and MySQL 5.6 is supported. I do encourage people for this sort of question to head over to GitHub as you are more likely to get people who built a section of the code to respond – better responses straight for the horse’s mouth. I run Windows myself on my laptop. My best suggestion is upgrade to the dev beta just dropped, wipe the var/ directory, and try again. There have been a few other people having problems with installation for various reasons – all have been resolved. Not many on Windows though. Oh, make sure you get the most recent version of Composer – there was a relatively recent change that makes it run much faster (a month or so back?).
Hi Alen, Thanks a lot for your reply. In fact I was able to resolve this little issue and installed long waited magento 2 on my Windows machine running php 5.6.3
Magento2 so far so nice and your blog is extremely helpful for good insights about magento2.
Thanks again for this wonderful Christmas gift, it’s my great privilege to talk with magento2 chief architect.