Author Topic: Concern with present architecture/platform  (Read 18197 times)

yed

  • Newbie
  • *
  • Posts: 16
    • View Profile
Concern with present architecture/platform
« on: June 06, 2007, 11:45:14 AM »
I was reading through the architecture in the OpenGoo pages. Part of which I have copied below.
Architecture:
"Current work is being done in Java/JEE, due mainly to the background of the development team.
We aim to develop a system that is:
  • Easy to install
  • Cross-platform
  • Secure
  • Open Source
  • Easily Extendable
"

I am not sure if the initial team has tried to look at some other viable options to raise the architecture on.
I think if its the above list of items then Ruby/Python may be a much better option than Java. I am not a java expert and neither an expert on Ruby/Python but I would say that Java now is like what C/C++ used to be to Java before wrt to Ruby/Python. I think, for example Python provides a much richer and neater experience for the developer than Java.

Also since a very large team of people will work across the globe on this product the idea would be to decide on a platform which is SIMPLE (yet powerful!) for people to start and continue. The simplicity of the system and platform will allow more people to join the project.

The problem with this architecture platform that is being setup is that it does not consider one very important aspect of a sound software system - MAINTENANCE
The platform and the architecture should be such that it is easy to maintain and easy to modify. The maintainability will ensure that people understand and pick up the work more easily at any stage of the development.

And another aspect which should be considered for OpenGoo would be to ensure that it is completely open source and is open source proof for the future.

By the way under what license is this product going to be released? GNU, MIT, BSD, Mozill etc etc..
And why?

A sound beginning will result in sound execution and finally a sound application.
These are just my suggestions and would greatly appreciate if people comment and add their views on it for OpenGoo.

Thanks.


conrado

  • Administrator
  • Hero Member
  • *****
  • Posts: 1000
  • Conrado
    • View Profile
    • Feng Office
    • Email
About the programming languages
« Reply #1 on: June 06, 2007, 03:12:22 PM »
Wow Yed, that was a very thoughtful comment.

In the older version of the site I had a poll asking which language should we focus on. Unfortunately, the poll did not get many votes. This is the one question that has kept rolling around my head since the first days of OpenGoo.

The languages/platforms that I mostly considered involving OpenGoo were:
  • PHP
  • Python/Plone
  • Ruby/Rails
  • Java/JBoss

Every language and platform has its advantages and disadvantages, and the arguments are never-ending. I have no experience with Ruby, although I would like to. All of the team members are very good with Java, so that is why the little programming we did was made in Java. But we are back in defining the mashup architecture, so no programming is being done at this time.

But, resuming the language issue, bottom line is OpenGoo will not have an official language. That is because the new architecture approach is to build a framework made mostly of standards sets, mashup components, and mashup tools. Each piece can be written in your language of choice. One crazy idea that even crossed my mind was to make various versions of OpenGoo:
  • The PHP OpenGoo
  • The JBoss OpenGoo
  • The Ruby OpenGoo
  • The Zimbra OpenGoo
  • The Plone OpenGoo
  • The Jive OpenGoo

And then you could mix all components, independent of the language. Of course this would take for ever, so right now we are just aiming at one version, selecting each component on its functionality merits, and not by language. Coincidentally, I think most components will be written in PHP, which seems to be the most popular language for OS projects. Note that by 'component' I refer to: 'the wiki' component, 'the blog' component, 'the spreadsheets' component, etc.

So, just like Google and Google Apps don't have an official language (I think Writely was written in .Net and Spreadsheets in Java), OpenGoo doesn't have one.
Get Official Support for your Feng Office. Support the development team. Sign up for a Free Trial here.

conrado

  • Administrator
  • Hero Member
  • *****
  • Posts: 1000
  • Conrado
    • View Profile
    • Feng Office
    • Email
About the licensing
« Reply #2 on: June 06, 2007, 03:20:47 PM »
Another tough question that I would have liked to decide by poll.

The 'spirit' is to give all products away in the most 'free' possible way. To be honest, I never completely read none of the licenses specs. I guess it would be one of GPL or BSD, or both. Suggestions are open.

Then there is the issue that we are going to use large chunks of other projects or, most probably, whole projects. I expect some minor troubles with that. For that matter, each project added to the mix should be contacted, asking them if there is any possible issue.
Get Official Support for your Feng Office. Support the development team. Sign up for a Free Trial here.

yed

  • Newbie
  • *
  • Posts: 16
    • View Profile
Re: Concern with present architecture/platform
« Reply #3 on: June 07, 2007, 11:50:25 AM »
I know its a tough choice when it comes to a project like this. But these are my observations and I might be just not very right. Most of the open source projects on the collaborative front have started with ground up having a unique flavour of their own (be it plone or zimbra). They would have used ideas from other systems no doubt but finally they are independent of each other.

The reason a mashed up environment might make things confusing would be because we don't have a common platform of agreement. The entire product will be split across the applications and each will have its own development cycle with none able to contribute to the other. The start might seem easier and look plausible but later on we might have to redo the whole thing again.

I would prefer something like the ZOHO guys have done. I "assume" they have a common platform on which finally the whole set of applications are running.

Doing what Google does might look easy but do we have that kind of dedicated manpower and resources to make it seamless?  Their philosophy is slightly different. They bought over the entire suite of online office and have given a seamles look to it with its existing set of applications.

I read about the other article about popfly from MS. Well I felt it was like Yahoo pipes.

I would like to start off with design of a simple engine which will be the backbone for the whole suite of applications. This engine would provide the common functionalities that each of these applications need. Much like a wiki engine (not exactly though) with a plugin like architecture for including the rest of the applications. These applications running on top of it will give the user the kind of look and feel which the application requires.

We can start of with only a rich text editor and spreadsheet before getting the rest of the suite up.

I am still not very convinced about the mash up idea right now. It sounds easy but then raising a family of kids from other different families together doesn't seem rosy.
 

conrado

  • Administrator
  • Hero Member
  • *****
  • Posts: 1000
  • Conrado
    • View Profile
    • Feng Office
    • Email
Re: Concern with present architecture/platform
« Reply #4 on: June 07, 2007, 03:14:40 PM »
That is the approach that first came to our minds. Then we decided to use JBoss as the platform. We are Java gurus (  :D ok, we know Java).

We could use the JBoss Portal and write portlets for each app. JBoss already has portlets for Wikis, Blogs and Forums.

Then... we might get very popular with the JBoss crew, but everybody else is going to seat there and stare, and all other platform's fanboys will be criticizing and expecting our failure.

We could have used Joomla as a framework. Even WordPress might be a good start (did anybody notice how 37signals' Highrise interface looks a lot like WordPress?)

Plone and Rails are other strong candidates.

We could write YAF (Yet Another Framework).

And then the idea of "the mashup" became slowly a good option. Google is doing it. Microsoft has now popfly (that is why Ignacio brought it up), Yahoo has pipes...

So, OpenGoo could attempt to bring one up. After a lot of thought, I started convincing myself that is not that hard, and there could be huge benefits: We would stand on the shoulders of other communities, having (indirectly, of course) much more support. We would empower other projects, instead of competing with them. Each piece could be easily replaced by a completely different app, over a completely different framework.

The OpenGoo.org site intends to be a proof of concept. At the time, the main page and the about sections are static html. I used this free template from Solucija to build it. Then I though maybe someone adapted this template for WordPress. Turns out someone did. Now if I get to adapt this template to the forum, the wiki, the spreadsheets... I think that is kind of what Google is doing (adding a couple of orders of complexity magnitude).

Next steps would be to have a single sign on and automate the template transformation to export a simple index, header, sidebar, footer and css to each app, and we already something that looks and acts quite integrated. Unifying contacts information would be next. And then... I don't know, my neurons are aching, plus I have a cold today.

I look forward for your input.
Get Official Support for your Feng Office. Support the development team. Sign up for a Free Trial here.

yed

  • Newbie
  • *
  • Posts: 16
    • View Profile
Re: Concern with present architecture/platform
« Reply #5 on: June 07, 2007, 04:54:23 PM »
I guess then the ploy should be to look at existing and pick the best of each and provide a way to run them all together under a single banner. - Is this what you mean?

Or better still take an existing open source project and build on it. (Say Zimbra)


conrado

  • Administrator
  • Hero Member
  • *****
  • Posts: 1000
  • Conrado
    • View Profile
    • Feng Office
    • Email
Re: Concern with present architecture/platform
« Reply #6 on: June 07, 2007, 05:08:42 PM »
I guess then the ploy should be to look at existing and pick the best of each and provide a way to run them all together under a single banner. - Is this what you mean?

Or better still take an existing open source project and build on it. (Say Zimbra)



Yes and... I don't know... I'm not sure. The second choice is almost joining the Zimbra team. It could give us a quicker start... But less communities would be involved. I believe that if we achieve the integration and inter-operation of a set of projects developed by different communities we could be greatly empowering the Open Source community in a way that it is not being done at the moment. Zimbra is a company, so I am not sure how many developers it could attract.

With the first approach, we will have all the developers of the other projects "working for opengoo". The sum, I think, will be greater. OpenGoo would provide "the glue" (by means of standards, tools, frameworks and a community) for them to integrate with other projects.

There is a topic created for the kick start of the first approach, suggesting some candidates for integration. The idea is to decide which ones will be going in the first milestone.
Get Official Support for your Feng Office. Support the development team. Sign up for a Free Trial here.

conrado

  • Administrator
  • Hero Member
  • *****
  • Posts: 1000
  • Conrado
    • View Profile
    • Feng Office
    • Email
Re: Concern with present architecture/platform
« Reply #7 on: June 15, 2007, 01:54:14 PM »
I found activecollab through the comments in this excellent post at Read/WriteWeb.

I think it is a better system to build upon than Zimbra, because of its simplicity. Also, the birth of activeCollab shares a lot with OpenGoo, so I feel both projects could work nicely together (The difference: activeCollab was inspired by 37signals' Basecamp, while OpenGoo by Google Apps).

Of course, this would mean marrying PHP in the mean time.

Any thoughts?
Get Official Support for your Feng Office. Support the development team. Sign up for a Free Trial here.

yed

  • Newbie
  • *
  • Posts: 16
    • View Profile
Re: Concern with present architecture/platform
« Reply #8 on: June 15, 2007, 03:24:15 PM »
I think there is a big difference in the philosophy of activecollab and opengoo.
Activecollab has married collaboration to a blog whereas opengoo is trying to implement a web office suite.

They are rather remarkably different. I dont think activecollab has any web office application in their software.

The things that look closer to our needs in activecollab is their plan to:
# build richer interface
# plugin support

We can plugin the opengoo apps to activecollab once opengoo is ready.

As for php like I said before as long as we are ready for the future concerns of opegoo anything is fine,  just that the app/product should be future ready and not be a old piece of code in future.

And finally I think opengoo is going to be a web office suite with blogging (like active collab) plugin to it.

yed

  • Newbie
  • *
  • Posts: 16
    • View Profile
Re: Concern with present architecture/platform
« Reply #9 on: June 15, 2007, 05:05:19 PM »
I tried Zimbra and Zoho on the web today.

Zoho is really what openGoo is trying to do though in a much more associated way.
I mean though zoho has a lot of packages/apps they are each isolated from each other. It would be nice if they can remove this separation and bring a single point of entry and view over the entire suite.

Zimbra is still very much behind Zoho in many areas. Their application seems to have started off as email client and now they plan to integrate the web office suite in it.

Try Zoho to get a feeling of the kind of features we can expect and what to ask for.

conrado

  • Administrator
  • Hero Member
  • *****
  • Posts: 1000
  • Conrado
    • View Profile
    • Feng Office
    • Email
Re: Concern with present architecture/platform
« Reply #10 on: June 20, 2007, 03:17:08 PM »
On activeCollab and OpenGoo

I am suggesting activeCollab.com as a framework, merely so we do not have to build one from scratch. I like their concept. It is a proven one: it is a copy of 37signals' Basecamp.

About 37signals
Why I like 37signals so much? I have tested all their products, and I love them. Read/WriteWeb had a poll about what online collaboration platform its readers were using. The results were impressive:
  • Google Apps 47% (743 votes)
  • Basecamp 18% (290 votes)
  • Zoho 10% (165 votes)
  • ThinkFree 8% (132 votes)
  • Zimbra 2% (33 votes)

So, besides the obvious leader (Google), I think we should be following Basecamp closely (and Zoho, ThinkFree and Zimbra; in that order). Starting with activeCollab and work from there could give us some advantage.

This doesn't mean we have to "marry" activeCollab. It is a one-man show, a great one, but that could pose some troubles, since Iljia Studen (the "one-man" whom I wish to congratulate) plans might conflict with ours.

About Zoho

I like what they are doing, and they are gaining much press. Their sponsorship at techmeme.com is paying out.  :P

As I said, a good company to keep an eye on.

About Zimbra
Yed, you are absolutely right. They started as an MS Exchange alternative. They are now realizing the real trend. And probably our only "competition". I wish them the best of lucks, since they are among the few of this truly embracing opensource.
Get Official Support for your Feng Office. Support the development team. Sign up for a Free Trial here.

Karl

  • Sr. Member
  • ****
  • Posts: 339
    • ICQ Messenger - 464979879
    • View Profile
Re: Concern with present architecture/platform
« Reply #11 on: December 29, 2008, 03:03:10 PM »
This was a very interesting thread and I learned a lot about the roots of OpenGoo. Reading this made me just more confident about OpenGoo then I was already.

 :) :) :)

setekh

  • Newbie
  • *
  • Posts: 35
    • View Profile
Re: Concern with present architecture/platform
« Reply #12 on: January 21, 2009, 05:56:21 PM »
I think that PHP/Java combination is the best for now.. Perl/Ruby/Python are superb and more powerful but other problems may arrive from using them or frameworks based on them, and you may expect much more community contrib. if you keep the app. as clean as possible. But that is what i think...

bayonian

  • Newbie
  • *
  • Posts: 32
    • View Profile
Re: Concern with present architecture/platform
« Reply #13 on: February 13, 2009, 12:43:17 AM »
Wow, this is where all the fight over the platform talk began. Interesting indeed.

Although, I'm new to the Open Source world, the projects like Drupal, Word Press, Joomla are so popular among the industry. They are all developed with PHP.

OpenGoo developers has chosen PHP is a right choice.


Pet

  • Guest
Re: Concern with present architecture/platform
« Reply #14 on: April 22, 2009, 04:36:20 PM »

About Zoho

I like what they are doing, and they are gaining much press. Their sponsorship at techmeme.com is paying out.  :P

As I said, a good company to keep an eye on.

I came across this post just now a bit late for the discussion, but I thought you might find it interesting to know that my company is a Zoho partner.  Because this will be publicly searchable I don't want to jeopardize my company's relationship with them, nor is it my intent to disparage Zoho. However, there have been a lot of mistakes that they have made in regards to their choices on how they run their business model. Everyone knows you can't be all things to all people, so you must choose.  They frequently release buggy versions of their software with known glitches. They want the public to know that they are very active and on the heels of google etc.  But this is at the expense of something else - support.  Their support is pretty bad.  In fact, even as a partner company, they do not give priority support to their partners.  They claim they do, but partners use the same level support as the public, which means their public support forums and phone number. They also claim to offer a return call within 24 hours, but in one instance we left numerous messages and they finally replied after 600 hours, nearly a month. This was for a critical problem that one of our customers was having.  Needless to say we no longer actively promote Zoho.

A lot of people like Zoho, and I do to some extent, their software is imho revolutionary, but their business practices leave a lot of people with broken or underperforming software with little or no recourse.

I recommend you learn from their mistakes and provide not only good support for your users, but also the highest level of priority support for your partners.

cheers

 

anything