Feng Forum
Other Topics => Development => Architecture => : bnordgren September 11, 2007, 02:18:08 PM
-
I really like what you're trying to do. I'm kind of starting from scratch here, so forgive me if what I'm about to say is a little naive...
Here's what happened when I first got to your site: I noticed that you "integrated all the features of ActiveCollab" with your first alpha release. Not knowing anything about activecollab, I clicked on the link and went to the site. First thing I noticed is that there is no information typically associated with an open source project on the activecollab page. No current release, no friendly "download" tab, no mailing lists, no bug tracker, not even a link to a comprehensive list of features. Instead, there's information on how to buy a license. :o If I arrived there from a google search instead of from your site, I wouldn't even know it was an open source project and I still don't know how to download it.
Fast forward through the next hour or so spent on the forum. No particular goal in mind other than to get a feel for what features the project has, what kind of support the development team offers, etc. Enter the realization that it's a one-man show. This, combined with the 0.7.1 version number that you mention, combined with the fact that a "summer 2007" release has been promised for a good long while and it's not here in spite of the fact that it's now Sep 2007....Well, my impression is that the one-man is either burning out or other priorities are asserting themselves. This is fine and it happens to everyone. It's just that when it happens to one individual on a large development team, the whole project doesn't suffer. (I was recently reading about the Apache Incubator and the reasons for not promoting "some code" to a full fledged project until a viable community had developed around it. Activecollab would seem to be a case study of why that is necessary.)
In any case, I did some more googling and I found this "project-open" project. (http://www.project-open.org/ (http://www.project-open.org/)) This website comforts me with the .org extension, the "download" link on the front page, the conspicuously advertised version of 3.2 indicating maturity and stability, the comprehensive array of modules which can be strung together however one would like.
At first blush, project open would seem to be a more secure foundation on which to build. I'm totally inexperienced with this "mashup" concept, having wasted my youth programming in C/C++/Java/python and avoiding both GUIs and the web. I suppose that it is possible that project open is somehow technically incompatible with this mashup concept.
Now I want to leave you with this other thing I found while googling for PM solutions. It's a good read and a really good, very lengthy, summary of open-source offerings. But the page, found at http://linas.org/linux/pm.html (http://linas.org/linux/pm.html) contains the following plea:
Note to Free Software Programmers: Please DO NOT create yet another project to implement some system! There are too many half-finished, half-functional systems already! Do some research, find a system that is appealing, and volunteer to carry it further! At first, this may seem less rewarding, because you won't be Mr. Super-Duper I-am-Head-of-the-Project Big-Shot Might-Be-Linus-Torvalds-Soul-Brother. But in the long run, you will find much more satisfaction from working on a project that everyone knows and respects, than being the sole author of and world-wide expert on some poopoo-kaka that no one has heard of. I speak from experience!
With this in mind, (and please realize that I am just trying to provoke thoughtful consideration and not to belittle your efforts): Have you considered the potential benefits of joining a project like "project-open" to integrate text and spreadsheet editors and craft the document repository to your liking? You're desparately going to need a community, and they come with one pre-installed.
-
bnordgren,
we - at OpenGoo - really appreciate you took the time to write all
this down. We consider it gives us a good chance to explain why we are
here.
The first thing I would like to state is that we chose ActiveCollab as
a starting point to build a Web Office, as oposed to building
yet-another-project-management-tool.
Our plan is to take advantage of the architecture, functionality and
features of ActiveCollab but not to be tied to it in any sense. As
just stated, our objective is different from the one of ActiveCollab.
You are right when you say that adding document edition to
project-open looks like a great idea. But what do you think the result
would be? I guess we would have an excellent project mangement tool
with document edition features.
We feel that this might be a good idea for some users, but that
approach is different from ours. We want to give users the possibility
to access Opengoo, create different types of documents, save them,
share them or even publish them. Project management features are not
our first priority at this moment, because we think there are lots of
them. However, we couldn't find any open source web office.
-
bnordgren,
Thank you for your thoughtful input. You brought up so many issues that I do not know where to start, but will try to address all. Note that I speak on my behalf, and other OpenGoo developers may not necessarily agree with my views.
On the use of AC as "the platform"
Marcos already explained the reasons for choosing ActiveCollab (AC) from the 'features set'-point of view. Also, it was written for the AMP stack (LAMP (http://en.wikipedia.org/wiki/LAMP_(software_bundle)) or WAMP (http://www.wampserver.com/en/)), making it very easy to install (another plus).
This choice wasn't easy. The time spent researching for the platform may prove insufficient. But there are so many possibilities that it is becoming virtually impossible to evaluate all of them. Then we kind of went with our guts.
The (or at least mine) biggest concern was exactly what you point out: the community behind it. We wanted a pre-installed community and, at the time, AC didn't 'seem' to present one. But since the issue was so big at its own forum, we figured:
* Someone would fork it (Already happened. Look at PP (http://www.projectpier.org/))
* We would fork it (Probably we won't need to, if PP meets our needs)
On Project-Open
Project-open looks very interesting. This is a topic of great interest to me. I never heard of them, and you could see I have taken a lot of time to research (http://theoserp.blogspot.com/) this. In particular, I am convinced that the future of ERP is on the Web (http://theoserp.blogspot.com/2007/05/on-web-office-and-erp-merging.html), and that such systems will succeed only when built on what Marc Andreessen just called a Level 3 Internet Platform (http://blog.pmarca.com/2007/09/the-three-kinds.html).
Building/Using a Level 3 Platform is one of the goals OpenGoo should pursue.
On the other hand, Project-Open looks complex, rather old and, yes, it is not developed for LAMP, which takes many points away from it.
On programming in C/C++/Java/python
You call programming in C/C++/Java/python wasting time?? Dude, that sounds like a lot of fun!
On the last link and quote
Lovely! Although a little old, the article you linked makes for a very thorough resource. And long! In fact, having taken the time to read it is part of the reason I took so long to answer you.
Now, by this time it should be clear that I totally agree with the paragraph you quote. But we did find two motives for making a new project:
1) There was no Open Source Web Office project that we were aware of.
2) There seems to be no relevant Open Source Level 3 Internet Platform for this kind of systems, written for LAMP.
Plus, you left out the immediately following paragraph on the article, which would be quite encouraging for us:
If you really want to be famous, then develop a core technology that allows one to easily create any one of these systems by merely modifying a configuration file. Most of these systems have broad similarities: they maintain lists of things that can be searched, categorized and updated. Only the labels on the columns are different. This is ideal for abstraction. Broadly speaking, workflow systems have this kind of configurability. But there are other ways to create the abstraction as well. Do it!
;)
Hope to read from you again. :)