Recently in Project Management Category

I vetted many of the project management tools that I've been looking at with teams at work and Redmine was the overwhelming winner. The one thing it was missing was a calendar, but we've found a makeshift way of dealing with this by inserting a static google calendar with reminders.

http://www.redmine.org/boards/1/topics/show/950

Still a work in progress but it seemed to work well enough.

Then came time to install Redmine on our production server instead of the test one. Installing Rails wasn't too hard as I found some good tutorials

http://wiki.rubyonrails.org/rails/pages/RailsOnCentos5

I had been running Redmine via lighttpd with fastcgi, but I wanted to go with Mongrel for this install as lighttpd was a bit problematic.

Good instructions on the Redmine wiki and another site
http://www.redmine.org/wiki/redmine/HowTo_run_Redmine_with_a_Mongrel_cluster
http://ubuntuforums.org/showthread.php?t=674598

Figuring out how to fire up Mongrel on boot was harder, but found that eventually too (wish I could remember where I found this one)

mkdir /etc/mongrel_cluster
ln -s /home/redmine/config/mongrel_cluster.yml /etc/mongrel_cluster/redmine.yml
cp /usr/lib/ruby/gems/1.8/gems/mongrel_cluster-1.0.5/resources/mongrel_cluster /etc/init.d/
chmod 755 /etc/init.d/mongrel_cluster
chkconfig mongrel_cluster on


Ok. Now the hard part. I knew that there would be some non-technical folks using the site, and I knew that I wanted at least logins secured with ssl. Problem is that Redmine runs under a subdomain, and in fact in this case it was a subdomain of a name based virtual host. I had tried once before to configure certificates to work in this case and failed. But I wanted to try again.

This time I won.

First I had to edit ssl.conf and add in the virtual host (as in the httpd.conf file) and also added this line to take name based virtual hosts
NameVirtualHost *:443

Then I edited openssl.cnf

Under both [usr_cert] AND [ v3_req ]
I added
subjectAltName = @alt_names
and
[alt_names]
DNS.1 = name1.com
DNS.2 = name2.com
DNS.3 = sub.name2.com

where name1.com is the owner's name as well (in the certificates, essentially the FQDN).

Then I found this post that took me through the rest (with the exception of the path to the certificates, which is in /etc/pki/tls in Centos/RHEL 5).

http://nixcraft.com/server-configuration-tutorials/3075-postfix-mail-server-create-self-signed-ssl-certificates-cent-os-redhat-linux.html

Did a little magic with the SVN repository as I had before, and voila it is up and running.

This week I had an occasion to start up a social networking site for some family. I wanted it to be a closed site where Internet newbies and parents of kids who are participating would feel safe. For those same newbies I wanted it to be easy to use.

I had previously looked at Elgg. But I decided to cross Elgg off my list for two reasons. First, it is transitioning to this part framework/part ready-to-run solution, and right now it isn't really either. Second, I think the "classic" version (which is the existing ready-to-run solution) serves as an educational networking space really well, but as a pure social space it fits less well.

I searched around and found several potential candidates. I narrowed it down to Insoshi, Lovdbyless and PeopleAggregator. There are many other solutions out there, but I wanted something that didn't have pluggable components that needed to be upgraded or any particular advanced tools that I don't think this community will need. I also wanted something that was under active development.

Insoshi and Lovdbyless both run on Ruby on Rails. I've been increasing my expertise in deploying RoR apps, but I'm still relatively new so these both took a while to install. PeopleAggregator is a standard LAMP app and took quite a bit less time to install.


Lovdbyless

Lovdbyless took the longest to install. It required quite a number of Ruby gems, not all of which installed in an entirely straightforward way. The biggest problem was rmagick, which is a Ruby library for ImageMagick. I run CentOS 5, which only has ImageMagick 6.2.x, which is not sufficient for the latest version of rmagick. Luckily I found this post on getting the previous release of rmagick to run. I also needed to install the ImageMagick-devel package for it to compile.

The instructions are pretty detailed for Lovdbyless, but they only get you running under WebBrick, which should be considered for testing purposes. In fact they brush over instructions for running this in any kind of production environment. The lack of details on getting this running under other web servers or in a production environment makes it somewhat frustrating. I got close with my existing knowledge, and I'm sure that a more experienced RoR sysadmin could get it going with Apache/mongrel or Lighttpd, but it shouldn't be that hard.

In the end the application was reasonable but pretty bare bones. Blogs and photos.



The administrator capabilities are limited to deleting users, and there was no way (at least in the GUI) to do anything else such as close or partition the community. In the end, I think that Lovdbyless is intended to be a framework for others to build on.


Insoshi

Insoshi is trying to do just that (build on Lovdbyless). They have developed their own social networking package, but have borrowed some pieces from Lovdbyless. It was a bit easier to install, though I still couldn't get it running just right under anything but WebBrick. Even then I had to jump through some hoops (also true of Lovdbyless) to get it running in production mode. I had to manually start the Ferret search server.

Insoshi shows a lot of promise. It has a good community for its age (which seems to be numbered in weeks or months). But it is even more of a bare bones app than Lovdbyless.




PeopleAggregator I'm still trying to get a handle on the story of PeopleAggregator by Broadband Mechanics, but it is LAMP app with a reasonable (though still not great) community that seems to support it. I found a wiki and some discussion threads for a few of the problems that I had during configuration.

Unlike the RoR apps, it got rolling pretty quickly. I did need one extra domxml php package that was available via yum, and in order for it to spawn extra subdomains (or sub-subdomains in my case) I needed a little DNS magic, but all of that was pretty well explained in the setup section.

PeopleAggregator has a much more robust set of tools. You create different networks (subdomains) that can be made private. Within those you can also create groups . Individual users can create posts of several types (blog, audio, video, photo) and with a simple checkbox they'll appear on the network's home page.



There is a pretty extensive configuration section, but some configurations need to be made in the local_settings file on the server. It would be nice if there were an admin account that configured all of those settings, but I couldn't find it. It did, however, make the first user I created have special configuration powers, but it still couldn't touch all of the settings that could be made by hand.

After some tweaking I was able to make a network private, and in fact I was able to make all networks private by editing some things by hand. That doesn't prevent people from creating accounts on the server, but they can't join any networks. It is notable that PeopleAggregator accepts IDs from other sources (Yahoo etc.) and it also will read status from Facebook, AIM and a few others.

It isn't perfect. But it is pretty good, and I like the ability to svn update to bring in any patches rather than having to download, and transfer settings. For now this is my choice. I'll see what the rest of the people say.

---

Redmine Revisited

I've briefly reviewed Redmine before. But I saw that they updated to a new version recently and in my search for a project management package that spanned technical and non-technical users I thought I'd give it another spin. It is another RoR app. But this one has quite good documentation on getting it installed and running under a number of different servers. That combined with my past experience had me up and running pretty quickly.

Redmine offers a nice suite of tools in a not too complicated package. With the improvements in this new version the usability is quite good. The one thing that I struggled with was integrating SVN. I just couldn't get it to connect. I posted in the Redmine forums (which use Redmine, which is a good sign) and had some suggestions, but nothing worked. But then another user in the forums solved it. With the SVN integration, the package really stands out. News to the front page of a project includes SVN updates and checkin comments, which is great.

The rest of the tools (Wiki, Files, News, Documents and the Forums mentioned above and shown below) are also pretty good.



There are better wikis and I miss being able to submit tickets by email (though there was word of this appearing in the next release). But it is above the minimum in just about every category that I'm looking for and could replace several apps (Wiki, Ticket Manager, SVN viewer, News) in one fell swoop.

Shortly after I posted about IceCore, I noticed that there was a VMWare image available on their website. That was the easiest way to get going with IceCore, so I fired it up on VMWare Fusion on my Mac. I did have some problems connecting at first. It turned out that the image was only listening to eth0 and my connection was being detected on eth1. I went to edit the network settings, but alas my favorite editor (nano) was not in the image. I limped through vi, and got the settings to work, and poof there was IceCore.

In terms of features, I think IceCore has the perfect (and I do mean perfect) mix. It is easy to create new groups ("Teams") of users to work on projects, and then provide them with a common set of tools that includes Blogs, Calendar, Discussion, Files, Milestones, Surveys, Tasks and a Wiki. This is great combination of tools



There is tons of ajaxy responsiveness in IceCore (note that the proprietary version goes by "Novel Teaming"). But the UI design needs some work. Threaded discussion take up a lot more space than they need to, and some of the features and layout will not be obvious to novice users.

These shortcomings sent me in search of related projects, when I found Liferay. At first, I couldn't tell the difference between Liferay and IceCore, but then Liferay did an update to version 5.0 and I was able to tell them apart. It turns out that IceCore is built on (the earlier version) of Liferay. Both have a similar portal display that allows you to add lots of widgets (both internal "portlets" as well as external sites and google gadgets).



It does add quite a number of features that facilitate working in teams, but much of the underlying functionality remains. There are numerous bundles of Liferay available, and I chose a bundle with Tomcat 5, which simply required me to run a start script. This uses hsqldb (which is fine for testing) and also installed a lot of sample data (which users on the forums have told me how to get rid of if I choose to continue with Liferay).

The new release of Liferay (which I assume IceCore will eventually incorporate) makes some welcome changes. For example, this is how threaded discussions appear in Liferay 5.



While this is how it appears in IceCore



Notice the popup frame that I'm sure seemed like a good idea at the time, but it isn't how much discussion forums work. New users like to have familiar functionality.

Similarly, this is how you see Wikis in Liferay.



And this is how you edit them.



These are pretty straightforward. But IceCore adds additional complexity with a Wiki that takes up more space, multiple views that are confusing and popup frames.



But where IceCore shines is in having all of the functionality that is needed in a one step install. Liferay is missing tasks most importantly. These could probably be added with a portlet, though I couldn't find one that was free and easy to integrate. IceCore also makes it really easy to manage teams. Liferay has similar functionality in Communities, but the Teams, along with the spaces that get created for them, are easier from an admin perspective.

I've seen some new UI designs for IceCore in their forums, and if they follow through on those and then build on the new Liferay, I think they'll have a real winner. Still, at this point in time IceCore is one of the best products I've seen. Speaking of which Chronopolys has gone 1.0 final and now includes an easy installer. I've also started looking at the community edition of Alfresco which has some new collaboration features in there. I'll try that next (after I can get it to install - it has been fighting with Liferay).

Recently a few collaborative software projects from blogs past got updated. I took the latest builds of OpenGoo (previous review) and eyeOS (mentioned but not reviewed previously) for a spin.

OpenGoo Take 2
The latest version of OpenGoo is vastly improved. OpenGoo is essentially activeCollab (the old open source edition now know as projectpier) with additional document handling. While it may sound minor, it actually is the one place where activeCollab/projectpier fall short, so this addition is quite welcome. The current build of OpenGoo (0.51 as of this writing) has had a lot of UI redesign with many AJAX additions. The result is a much more cohesive set of applications that no longer feel cobbled together. Instead, the design (including a new integrated menu) makes this feel like a top notch product. Combined with the great project management features already in activeCollab/projectpier, this makes for a very viable product. It is a simple and typical LAMP install.



My one concern about OpenGoo right now is the potentially small developer and user community. While it is a highly active project on Sourceforge, there are few if any postings in the forums on the website or on Sourceforge, and the website itself is sparsely updated. I hope this is because they're putting all of their time into development.

EyeOS in Short
The notion of a Web OS is intriguing to me. So far I don't think any products are really there, but the best I've seen is EyeOS, a WebOS you can install yourself.



The idea behind a Web OS is that you can log into a server and have access to all of your typical applications and data, while nothing has to run locally. EyeOS does a reasonable job at this, replicating somewhat of the look and feel of a desktop OS. What would be great is if you could actually access data on the server through the application, but alas it only accesses data from within the Web OS (the next version will allegedly offer a sync application to sync local files to the server). There are some decent applications - basic office apps, meebo for IM, etc. I haven't quite figured out the web browser yet, since you obviously have access to a web browser if you are using Eye OS. At first I thought it might be for privacy, but all of the cookies and cache are still stored locally. Still it has some potential.

Looking Ahead to Icecore
I recently started exploring another collaborative application called IceCore. From what I can gather IceCore was recently acquired by Novell, which bodes well for its future (in terms of resources anyway).

I haven't tried installing the open source version yet, but I tried out the demo online. There are some confusing things about the interface, but it has a great featureset, and once I got the paradigm figured out it was pretty easy and quite flexible.

I'll offer up a more complete review after I try to install it and run it myself.

Deki Wiki just went to a 1.9a release. And finally reverse proxying works on the standard Vmware image install. A quick install, IP configuration and simple apache reverse proxying configuration (see the bug report) and it was off and running. I renamed the server, which meant further email configuration, but that was it. With Deki Wiki installed I took the chance to play around with it a little. It comes with some very slick skins.



Editing pages via WYSIWYG is very easy. The new version has preliminary support for FCKeditor and TinyMCE in addition to the default editor. The default editor still has the most support for features, and the fewest bugs, but it doesn't work with Safari. I tried the FCKeditor, and it worked fine. The editor floats on the web page, which causes some problems, but it is just experimental for now.

Editing pages, creating links and uploading media is straightforward. I'd put Deki Wiki as my clear choice for Wikis at this point.

Redmine

I've been searching for a project management/trouble ticket system. So far I'm voting for Eventum, which I have configured to manage bugs and features. It accepts incoming requests via email, which is great and it was quite easy to configure. It does one thing and it does it well. But I've been hoping I might find a system that can handle tickets as well as a few additional project management features. Trac is a popular choice, and I'm going to evaluate that further. Redmine is a similar Ruby-based tool, with support for some additional features and multiple projects. I added it to my previous Ruby install following the instructions. Configuration was easy except I never got the external svn repository working. I'm not sure if it was just me or a more fundamental bug. Otherwise, the tool was quite well set up.



But in the end the features of the issue tracking weren't as good as Eventum, and the other features (Documents, Files, Milestones) weren't as good as the best of breed for these tasks. I'd rather have great issue tracking and handle the project management features elsewhere.

The Xwiki project has spawned a number of related projects. One of the recent offspring is Chronpolys, a wiki-based project management site. It still doesn't have a 1.0 release, but it is in release candidate phase. It felt a bit more like a beta, but it had a lot of promise.

It calls itself wiki-based project management, but the wiki is really just one part among many. It combines a solid feature set including timelines, versioned documents, email notifications and of course a wiki. It has an attractive Ajax-full interface though it suffers from some organizational challenges and some interface clutter. Much of the success of Project Pier (or its predecessor ActiveCollab or paid counterpart BaseCamp) may be attributed to the simplicity of their interfaces. Chronopolys has them all beat on feature set, and the interface is slick, but it needs some simplification.

Chronopolys comes in a self-contained zip package that allows you to launch it on Windows, Linux or Mac OS by simply launching a script. This made it great for testing. After a quick configuration you can access the overview/dashboard.



Which can be viewed by timeline (above) or by asset type.



From there you can delve into the details of an individual project. This is presented in a useful fashion at first glance. But the interface for adding or manipulating resources is confusing.

It tries to strike a balance between displaying all of the information about your project and keeping a clean interface. But as the number of assets grow the interface begins to get cluttered and could be challenging for users to find the information that they need.

The wiki is based on xwiki, which is really quite good. And in fact having a wiki as a central part of a project management suite seems like a good idea.



Mindquarry tried that, and unfortunately failed. I hope that Chronopolys can succeed. Most of what it needs is a little less rather than a little more. A little refinement and documentation and this could be a great product.

---

Shortly after my recent Elgg review, Elgg announced that they are starting a new codebase and forking the current codebase into Elgg Classic. This new version will be more modular and they will package several configurations for different purposes. The only down side seems to the that setting up a Elgg-based site in the short term is a bit dicey, though they claim the data is portable. The long view is quite positive and worth keeping an eye on.

After my disappointment with activeCollab's switch to a purely commercial model, I found OpenGoo, which was built on activeCollab 0.7.1 and now will be built on the activeCollab fork ProjectPier. The mission of OpenGoo seems ambitious, to take on the Google platform - docs, collaboration, etc. They're trying to build up from existing tools and glue them together. I think that can work to an extent, but it does put them in the precarious position of relying on many other projects, along with their updates, security, and timelines. The case of activeCollab eludes to this vulnerability, but I guess it also shows that there is some resilience in that they are adapting to another project.

At this point OpenGoo is essentially activeCollab 0.7.1 with several FCKeditor based document editors including a word processor, spreadsheet and presentation system. While this may sound mundane, it actually makes up for one of the most serious deficiencies in activeCollab, the ability to generate documents from within. Basecamp has Writeboards, and activeCollab needed something equivalent. In many ways, these tools could advance the project well beyond Basecamp.

The word processor is the most mature of the applications. It is really just an FCKeditor form, and that works quite well. The spreadsheet is very rudimentary, and the presentation system (see below) while ambitious is really just a proof of concept. I can't see actually using this tool to make a presentation, and there doesn't seem to be a way to upload Powerpoint documents and edit them in OpenGoo, or conversely download existing documents as Powerpoint. The same goes for the other editors. This will need to be addressed at some point to become a viable option.



OpenGoo has a long way to go. I had a number of difficulties getting started. First, FCKeditor is set to use asp by default. Given that activeCollab is in PHP, it seems odd to leave this as the default. It took some time to diagnose this as the source of the problem (OK, it should have been obvious when I started getting error messages about asp scripts on my Linux server). Further modifications were required to allow FCKeditor to upload files, and additional directory permissions needed to be set to allow files to be saved (initially when I tried to save files I got a message that I later translated as the Spanish equivalent to "document saved" but it didn't actually save anything).

What OpenGoo really needs is some testing and documentation, but it shows a lot of promise. If they can get out some of the rough edges, while it may not replace Google docs, it could be a viable project management and collaboration tool in the near future.

I've been waiting for the update to activeCollab. The last beta (alpha) of the "then free" project was a solid alternative to Basecamp. They've been undergoing closed revisions and modifications for quite some time, and it was clear that they were moving at least in part to a paid business model. However, it seems from the latest pricing information that activeCollab has moved entirely to a paid model (or will move when it is finally released). That is unfortunate. I'm all for having paid options, but having a free alternative for "non-commercial" use would have been great. Perhaps it will still happen.

I did see in the comments, that ProjectPier seems to have forked the open source activeCollab and continues to develop that option. I also happened to stumble upon OpenGoo which was using activeCollab (and now will use ProjectPier). OpenGoo looks to attempt to be an open source version of Google Docs and other services. That will be interesting to keep an eye on.

Since my last post (in addition to working on a book for my other job) I've been hammering away on deployments of Zimbra. After quite a bit of testing I decided I liked it enough to deploy the Network Edition (paid) at work for a small workgroup, and at the open source edition at home for my family. One thing that didn't become apparent until I actually started working on these servers in production was that the default spam settings were a little weak. Fortunately there was additional documentation on the Zimbra wiki, as well as user contributions in the forums and on the wiki. I've found that there are a few rough spots (e.g. switching from http only mode to https mode broke smtp authentication). But the forums are active (although occasionally contentious) and help abounds. The open source edition is quite comprehensive, and other than lacking syncing, rendering of attachments in html (which is actually quite handy on the road), mobile capabilities and some scaling capabilities, it matches the paid version quite closely.

I've also been putting ActiveCollab through its paces on one of our projects and users have been pleased.

---

While I haven't had the chance to do some of my comprehensive evaluations on Wikis and Collanos (next on my to do), I have had a chance to tinker with some hosted tools that have been receiving a fair bit of press. Here are some quick overviews of a few of those tools.

Wetpaint
I am definitely in the market for a more usable wiki. I'm not sure why there is this anti-WYSIWYG bias in the wiki community, but it seems to be there. However, with wikis headed for the enterprise (for example see the Lotusphere announcements from last week) I think they're going to have to become a lot more user friendly. Wetpaint has to be at the top of the user friendly list. Wetpaint is a wiki designed for individuals, workgroups and small clubs/organizations. It is simple and slick. It is very Web 2.0, but it wears it well. At its heart is a simple ability to create and edit pages.



The edit this page link is always right there and brings up a simple ajaxy editing toolbar. That in and of itself is a reason to think about using wetpaint. But it gets a lot of other things right too. Adding a new page is done by... hitting the "add new page" button. And when you add that page it appears in a navigation menu so that you don't lose it. To link to that page you can easily find it using the link tool in the editing toolbar. This is the way wikis should work.

It does go a little overboard with the use of widgets, but I can imagine that these will evolve and could prove to be quite useful.



Wetpaint also easily facilitates sharing and commenting with others. You can choose to keep your wiki public or private, and weetpaint is free and ad supported. I have not found any ads to be obtrusive, in fact on my sites I haven't seen them at all. Wetpaint seems hard to beat, but I'll be trying to check out DekiWiki soon.


Koral
Koral is another Web 2.0 site. Its mission is sharing of documents, and again it does it simply and well in Web 2.0 style.



It is a well laid out site for uploading and sharing documents. You upload and publish content that you share with groups of people divided up into different workspaces. The interface makes it easy to get documents online, choose with whom you want to share them, and keep it all organized. Its one flaw may be that it is solely document focused. This likely means it would have to be used as one tool among many. This isn't necessarily bad, it means that there isn't a lot of feature creep or confusion about how to do thousands of different things. Perhaps it could be paired well with Wrike (see below). Right now Koral is hosted and free.


Wrike
I have to laugh just a bit every time I look at the Wrike page. Not because it isn't a good tool, but because it buys in a bit too much to the Web 2.0 thing. If you've read some of the sites spoofing Web 2.0 they'll tell you what you really need to do to be Web 2.0 is use certain colors, use particular fonts, make sure your logo has a reflection, and be sure to label it "beta". Well Wrike complies with all of these.



But beyond that it seems like the start of an interesting product. Wrike is a project management tool focused on tasks. Users create tasks and assign and notify others about these tasks. These tasks can be managed via a combination of web interface and the user's own email. The latter is what makes Wrike unique. It facilitates emailing in of information as well as notification out about tasks. Once tasks are initiated users can add documents, comments and updates to them in a relatively easy fashion. I have to admit that I haven't tried Wrike first hand, as the couple of times that I tried to create an account, it has been "down". Presumably they're in rapid update mode, and I will give it a whirl when they let me (or when it is out of beta). Right now Wrike is free.

I've been tinkering with activeCollab for quite a while now, but hadn't got around to using it enough for a full review.  This past week I needed to start a new project site, and thought that would be a good time to try it out in earnest. 

For those of you who don't know activeCollab, it is a near clone of the popular project management system Basecamp.   Basecamp, made my 37 Signals (home of Ruby on Rails), has an avid following, and API for add-ons and a free trial version.  The free version lets you set up a single project with unlimited users, but requires you to have any uploaded files stored elsewhere on a server.  This really only makes sense if you have a server in the first place.  Plans go up from there to $12 per month for 3 projects and 200mb of storage all the way up to $149 per month for unlimited projects and 20gb of storage.

Basecamp's popularity is likely due to its simplicity.  It is easy to navigate around, see updated information and do what you need to do. It doesn't have every bell and whistle but what it does, it does well and does with elegance.  This is critical for small collaborative projects.  If the collaboration software is not easy to use, people will not use it and either default to email (which is a pain) or cause the group to be dysfunctional.

Basecamp is about sharing messages, files, and milestones/to dos.   The same could be said for the PHP based activeCollab.  It is simple, elegant (just a little less flash than Basecamp) and focused on a doing a few things well.  Though it has a few shortcomings (read on) it is almost as good as Basecamp for zero cost (if you have a server). Others have discussed the differences between activeCollab and Basecamp, and I'll build on that as well.

activeCollab (I tested version 0.71) is a fairly standard PHP/MySQL application.  You unzip the archive into a directory, setup a database user and database and point your browser at the install script.  There were a few hiccups in the installation process including a misconfigured directory, write permissions and a reference to a missing "php gd" extension.  None of these were hard to fix (the php gd extension was easily retrieved through yum on FC5), but it would have been helpful to have a suggested fix for each problem rather than just stating that you have the problem. 

After installation you can log into activeCollab as the administrator and get started.  It defaults to a login screen with suggested first steps including adding users, projects and clients.  That is helpful, and it does keep track of your progress, though I would have liked a little less emphasis on clients, since we're using this product internally, as I suspect others are.  The notion of clients should feel more optional.

Each user has a Dashboard that provides the information in a quick view about all of the latest information on their projects.

From there you can quickly access your tasks or projects and narrow down the amount of information you are looking at.  The "most recent" feature is pervasive and one of the things that makes activeCollab so usable.  You can jump back into a project and catch up where you left off.  

Within each project the main tabs are

• Messages
  
• Tasks
• Milestones
• Files
• Tags
• Forms
• People

In my previous experience with Basecamp, Messages seem to be the backbone of this software.  From here you can post messages like you would on a forum, assign it to milestones, and provided several other specifications, most importantly who you would like the message emailed to. While web based collaborative systems are great, there needs to be a balance of push and pull.  I find that people need reminders of activity.  They can ignore the details and find them later online or read them from their inboxes.  That is really useful.

Tasks are assigned to individual users (or groups of users) and are easily organized and prioritized. They only get a sense of time if they are associated with milestones, which can be done easily.  Milestones are time based, and can be associated with tasks or messages.  In many ways Milestones are the organizing principle of activeCollab, but they can be a little less so since activeCollab (unlike Basecamp) also provides the facility to associate Tags with all items.  Milestones can be exported as ical files, and many parts of activeCollab can be subscribed to as RSS feeds.  Files can be uploaded and placed in folders that the user defines.  There is also provisioning for versioning of files, which is incredibly useful. I'm not quite sure the function of Forms, but you can create a simple one text field form that can be displayed and the response to which will be added to another section of the site.  I suppose the utility of this is if you really did have a client who only had access to very limited parts of the site, you could use this to get information from them.

So what is missing?  The main thing that is missing from my perspective is some sort of shared writing space.  In Basecamp this is their simple one page Writeboards, though this could be a Wiki.  As you can see in my screenshots, there is indeed a Wiki tab, but this is only there because I put it there (the joys of open source).  I made a three line change in one file that put an extra tab there that simply provides a link to a totally separate Wiki (I chose pmWiki, though this will need further evaluation).  It is really just there for convenience.  It would be great to have a Wiki (or Writeboard) included, as it fills a useful space between messages and files.  There is quite a bit of discussion on the activeCollab forums about this topic, and I'm sure at some point it will be included.  The disadvantage of a separate space is of course the complexity of multiple logins, lack of notifications, and the inability to integrate with the Dashboard.  

After some time stepping through activeCollab, it came time to make a decision about what software to use for my new project.  I went through many other possible choices (thinks like egroupware), but most of them are just much too big.  I want to give the team the minimum set of useful tools to make it easy to use (and therefore increase the likelihood that it will be used).  So I was deciding between activeCollab, the recently reviewed Epiware, and Basecamp itself. 

As far as activeCollab was concerned I saw the following:

Pros

• Easy and familiar (to anyone who has used Basecamp) interface
• Does most things that I want it to
• Great centralization of information and updates
• Great messaging features out to email

Cons

• No Wiki or writeboards
• Weak calendaring (milestones only)
• Developed by a small team (mostly one person it seems - slow but steady progress)
• Some minor UI inconsistencies (e.g. what happens when you click on the little down arrow?)
  
• Still beta and probably will be for a while

Epiware had Wikis and calendaring going for it.  But it lost on familiarity of the interface (and a little on ease of use) and had a number of annoying (albeit not showstopping) bugs.  It had enough bugs (problems with timezones, a few exceptions) that, although I felt like it has a superior feature set, I worried that for this project those bugs might annoy people enough that they either wouldn't use the software or would force me to fix them myself.  Perhaps my next project I'll choose Epiware as I really like the features, but for now it lost out. 

As for Basecamp, the Writeboards were the main thing it had going for it.  The free account would force me to use my own storage space anyway, which would mean going to the next price up.  While not terribly expensive, it wasn't clear what I got for that cost.  

In the end I went with activeCollab.  As we start to use it, I'll try to update this space.