Future and roadmap of Transfuse

[Alexey Frishman]

Hi John!

First of all, I'm so impressed with the library, it's realy fantastic what you are doing! I want to use it in my application, because it gives everything I need combined in one library: powerful DI, components, views, listeners, event bus.

My only concern would be that this library is less known/used, than Android Annotations and Dagger. Why is it so? Also it seems that you are the only developer, it's a risk. What are your plans for the future? Do you have any roadmap? What are you going to add/improve? Will you support this lib if some new Android version arrives, e.g. Android L?

Anyway, I'm so impressed that I must give it a try! Would be great to hear your opinion on these topics.

[John Ericksen]

Alexey,

Sorry for the delay getting back to you.  I saw this in my inbox but failed to notice it coming from Google groups (thought it was from github).  By the way, was this you: http://stackoverflow.com/questions/20306723/how-does-transfuse-compare-with-dagger#comment-38841997 ?

Thanks for your kind words.  I've (Really "we" because there are a number of people that have influenced and worked on Transfuse) put a ton of work into Transfuse and it's always good to hear it making an impact.

Yes, Transfuse is less known than AA and Dagger, probably due to the following reasons:

1. I'm not aggressively advertising it at conferences and such.
   
2. Transfuse isn't officially backed by a company.  I started Transfuse because I had an idea about how Android applications could be better in a way that was not developed yet (Roboguice and Android Annotations were the popular projects when I started Transfuse) and I followed that dream.  What you see today is purely the result of my passion for it and open source... Transfuse is not my day job, unfortunately.
   
3. Transfuse wasn't born out of Google like Dagger.
   
4. Transfuse is a bit harder to pick up because you really need a deep understanding of Android to appreciate it.  It's not for noobies, and I think that may hurt adoption.
   

For the roadmap, we have a good backlog of issues on Github to fix up.  I'm currently reworking the core of Transfuse around how it generates classes and how other Projects will plug into a Transfuse app.  After these are resolved and everyone is happy we're going to release 0.3.0.  I'd say this is then a candidate for a full 1.0.0 release, which has been a long time in the making and represents a stable API and functionality.  After these releases I think its just a matter of fine tuning Transfuse and making sure it supports the latest and greatest.  Yes, Android L is going to be supported.  If you're looking to get involved, there are a number of opportunities including working on the codebase (check out the issues), working on the documentation (all forkable on Github, check out the transfuse-jeykll-site branch) and just giving feedback on your experience.

If you're worried about me dropping Transfuse, check out the history.  I've been actively working on it for nearly 3 years now.  Plus, it's open source.  If I was to be hit by a bus the code would still be available for other people to use and modify.

Along the lines of a roadmap, there is also the possibility for side projects from Transfuse.  Parceler (https://github.com/johncarl81/parceler) was originally part of Transfuse, it was useful enough to warrant it's own project, so I broke it into it's own repository.  Now it's more popular that Transfuse.  There are certainly other candidates for this sort of thing.

I encourage you to give it a try.  There's a getting started guide on the website to get your up and running quickly (http://androidtransfuse.org/getting_started.html) and if you hit a snag don't hesitate to reach out here in Google Groups, on Github or on Stackoverflow.

Thanks again.

John

[Alexey Frishman]

No problem with delay, John. And thanks for your really in depth response.
Yes, it was me on StackOverflow :) I did not know, where to ask you first.

I have to find free time and learn more about Transfuse, there is so much to learn there.

BTW, are there any limitations of Transfuse you can think of compared to direct usage of Activities and Fragments? Does it cover the complete lifecycle? Any performance limitations?

[John Ericksen]

To the contrary, Transfuse is extremely performant across the board.  I have not yet put together a thorough benchmark but the difference between it and other solutions that use reflection is night-and-day.

Transfuse really shifts the work of analysis of your application to compile time, holding the runtime of your application, especially startup, sacred.