1. It looks awesome in terms of UI 2. The webpage looks great and explains everything I wanted to know.
But when I looked at the source code and saw Java I was deeply disappointed.
I use Java a lot, MapReduce, Pig, Storm topologies, I love java but I would never use it on a server.
With every open source project adoption is super important, having Java on the server is a big ask when you want someone to try your project out.
One thing that will help a lot here is if you have a Dockerfile on your repo, this way people can build and deploy it easily. I'd even go as far as creation a chef recipe for it. Those can seem out of scope, but can help a lot.
Anyway, this seems like a solid piece of open source. Awesome
I admit that I still don't understand why java receive so much hate.
What kind of language would you prefer for a server side application and why?
For this project I've chosen java _because_ it's a boring language, I really don't want to be enabled to write clever code that will be unreadable after few months.
Additionally, with a jvm I've got extremely good libs/tools for developing the application, and when in production a whole heap of monitoring tools.
So uh, sorry for liking the jvm :(.
(And it's not that I know only java, far from it...)
edit: and still, I don't think that requiring a jvm runtime is asking too much. I mean, if you use a managed language, you will need a runtime.
- lack of generics (sorry, but even though I kinda like go as a language there are features that I consider mandatory)
- lack of decent dependency handling (I don't consider vendoring a good solution, with maven central I know that the libraries will be there forever and there is a versioning scheme, pointing to the master branch of a git repo is not a good solution either, I value stability)
- lack of officially supported sql drivers and libraries like lucene (sorry but I don't really want to depend on a random github repo, it does not give me enough ensurance that it will be mantained)
(And I don't think that a statically linked binary/smaller binary/less memory footprint is a good enough reason to dump the jvm and his ecosystem)
https://github.com/digitalfondue/lavagna-docker
As you see it's pretty easy to switch from this "demo" Dockerfile to another one that uses a linked MySQL container. This repo is linked directly to Docker Hub:
You can follow the guide for openshift here: http://lavagna.io/help/openshift/ .
I'll see if the free heroku tier work decently and provide a guide too.
(And yes, it will work on a 5usd digitalocean instance, as it's strictly more powerfull than the free gear from openshift)
Also, is the implication that the merit of the tool is decided based on your personal liking/disliking for the language it is implemented in? That seems childish! If you are responsible for any decent business, dangerous!
- a .war for deploying in a standard 3.0 servlet container
- a self contained executable war that has a embedded jetty http server
For the self contained option, you only need a working java 7 runtime. I don't think that I can do it simpler than that.
Additionally, if you prefer an even more automated way of updating the software, the application is available on maven central, so you could (as we are currently doing it in our openshift instance) use the maven jetty plugin and select the depedency.
I can't really think of a simpler deploy system. (except statically linked binary, obviously, but there you need to select the correct os :))
edit: we have a docker image for testing purpose too.
I can't imagine why a company wouldn't get a proper yet slightly more expensive server with an adequate amount of RAM instead of a vhost that only allows to run PHP, especially as the software itself comes for free and most companies such a tool would be useful for likely already have some sort of dedicated server - either on-prem or hosted - anyway.
Besides, Java often is the only option for deploying tools in larger companies.
I have one question, though: What's your business model with Lavagna? From what I've seen at a cursory glimpse there doesn't seem to be dual-licensing or anything similar (which on the other hand probably wouldn't make sense for a project management tool anyway).
In fact we (3 of us) are all working in a "big" enterprise settings so we kinda know (and suffer sometimes) about the general constraints of deploying external software (it's more probable that a java or dotNet app will be accepted than something built with an exotic stack that require others additional software that nobody is able to operate internally).
Anyway, for running lavagna you don't need a big server, I'm running it on a small (free) 512mb of ram server from openshift _with_ a mysql instance in it! This is a 5usd/month instance at digitalocean.
>What's your business model with Lavagna?
We are developing it in our spare time (I don't work full time so I can do projects like that), so nothing concrete at the moment, we are open for any opportunity.
./lavagna.sh: line 54: -Ddatasource.dialect=MYSQL: command not found ./lavagna.sh: line 55: -Ddatasource.url=jdbc:mysql://localhost:3306/lavagna: No such file or directory ./lavagna.sh: line 61: -Ddatasource.username=lavagnauser: command not found
It does not use _that_ much of ram. (And yes, it has a basic trello import functionality).
Is i18n on the roadmap?
(Probably there are some issues in the gui for RTL languages and for CJK in the full text search support)
