Scale is the New Black

Cross-posted from the Particls blog.

bupropion sr in uk

Over the past two weeks, the Faraday Media development team have been hard at work migrating all our products and initiatives into a new datacenter.  The new data center was much more suited to Faraday Media technology – much easier to scale, much faster and more reliable.

As part of our efforts, we’ve finally had the opportunity to give Particls and the Engagd platform their own dedicated servers, effectively quadrupling our processing capability.  This will allow us to service our partners and customers with increased reliability and confidence.

We’ve taken the opportunity with the new servers, to finally move our blogs off Blogger and onto a hostedWordPress solution, giving us far more flexibility with our blogs and presentation.

One problem, however, after we upgraded to WordPress 2.6, was that when we changed the permalink settings (to something more tollerable than ‘?p=x’) suddenly, index.php worked fine, but any other page reported as ‘not found’.  After serverl long hours Googling for the answer, there was lots of “answers” for WordPress on Apache (specifically about correct access to the .htaccess file and ensuring the correct PHP/Apache modules are installed) – but none about how to solve these issues on IIS.  It turns out that there are a number of known issues with 2.6 on IIS, which are now solved with the release of WordPress 2.6.1.

With most of our migration issues now sorted, we can now confidently continue to deliver our attention and data portablity solutions to the masses secure in the knowledge that our services are scalable and our bandwidth is plenty.

Twitter “Track” is back – Introducing the Particls Fountain

Cross-posted from the Particls blog.

From the day we launched Particls 1.0, people have always been excited with our approach to Attention Management.  And while we certainly don’t consider it to be a failure, we always felt that being a Windows only desktop client, with some complex UI challenges, that there was something too difficult for many users to grasp.  There was effectively three paradigm shifts (Aggregate Everything, Rank against interests, Variable interruption based on relevance) and this was a too much for a lot of users to wrap their heads around – ultimately creating a ‘barrier to entry’.

We switched over to a web-based solution, cutting Particls in half.  The Attention Management Engine was removed and eventually became Engagd, and the visualization and alerting engine became what we called ‘Particls 2.0’.  We’ve been working on these two platforms for quite some time, but as the economy turned, and faced with ever increasing costs and minimal resources; we decided to find another way – to work with fine scalpels instead of axes like we once did.  A limitations of resources always forces companies to focus on what really matters – and we choose to use this economic downturn as an opportunity instead of allowing it to consume us.

In order to reduce complexity and scope, we’ve diverted all resources onto a new project we’ve been internally calling “Particls Fountain”.

Particls Fountain will eventually become what we wanted Particls 2.0 to be, but rather than building from the bottom up, we’re building it from left-to-right. Right now it is simply a replacement for the Twitter Tracking service, where you follow topics of interest you define and Particls alerts you Tweets about that topic.

Currently these alerts are delivered via XMPP or Direct Messages, but other mechanisms have been requested and are in the pipeline. Unfortunately however, Direct Messages are being limited by the Twitter API. We will be bringing new channels online to compensate. If you want to get started with Particls, simply follow the instructions at http://blog.particls.com/index.php/instructions

Aside from extremely agile development and releasing frequent, smaller updates to the service, we are also letting the community be the primary driver of development.  We’ve setup a UserVoice site where great ideas are already flowing from the community of about 100 testers.  This feedback is vital, and it’s encouraging to see these users vote for their favorite feature.  Its quite insightful, and it clearly demonstrates what we think is a cool feature, is actually not always what users want or care about.

As a developer I also find it extremely rewarding to mark a feature as “complete” and getting immediate feedback about it.  Its great and so far we’ve found that not only do we as a team produce code faster, but we also building stuff better than we did without it.

It’s still early days for Particls Fountain, but we really do want to make this a tool everyone will find useful, so please come try it out and give us your thoughts.  Be our bosses and tell us what to do to make it something you love.

Because we do.

Note: If you want to get started with Particls, simply follow the instructions at http://blog.particls.com/index.php/instructions

Braxton Hicks Contractions

Cross-posted from the Particls blog.

Stress levels have been pretty high for a few days at Touchstone HQ. We’re at the business end of alpha and experiencing the typical Braxton Hicks that a start-up experiences as they prepare for beta.

It’s been long tiring week and the team are exhausted. But it’s ok – it lets us know we’re still alive.

A (seriously) lucky few have had the opportunity to actually play with the beta version and so far the feedback has been extremely positive (perhaps they will write something pleasant in the comments to this blog post to make us feel special). The only problem preparing for the long anticipated beta is that, at some point we have to turn off the ideas (something we’re all struggling with).If you’re a newbie to our humble little blog, please make sure you sign-up (on the side <– over there) as a matter of urgency. Trust me! It’ll be worth it.

Beta

Cross-posted from the Particls blog.

Today (hopefully) will mark a milestone in the life of Touchstone. While Chris is in the States, talking with many of the people we’ve been engaged in conversation with since we started, I have been passed the helm of our Australian Operation in his absence, and prepping a build for tonight, which I *hope* will become our first Beta build.

With the Touchstone Team growing, we have surrounded ourselves in positive, interesting people passionate about what Touchstone is all about. This positive drive, has been accelerating the momentum of Touchstone and has given us the ability to produce a build which is far superior then anything we have previously produced.

I thought it might be interesting to post a few of the recent enhancements this build of Touchstone will bring, a sort of Elaborate Release notes to show you our direction and to convince those who have not yet signed up to the mailing list to receive this build, to do so now (over there <– on the left). So, here are a few of the enhancements of Touchstone 0.4.397:

The first and most obvious change is the new slick interface as it was just too difficult and too cumbersome to use the previous one. We designed this window to be a little more “normal” while at the same time empowering us to do more with it. Now all inputs, prioritization and interuptions are all controlled within a single window.

The user now has a more verbose status area, where they can view the items Touchstone is “tracking”, sources found, Inputs and Outputs. It also shows the current state of the automatic profiler (another overhauled system for this build).

The event log is another useful tool we’ve included in this build. Originally designed as a debug tool, we thought a way for users to see exactly what’s going on inside the core during Touchstone’s standard operation would be helpful. Especially for our Adapter Developers.

While the new interface still has a few things we need to iron out, overall it makes a massive difference to the Touchstone user experience.

Another big shift is the way ranks and ranges are set. Gone are the Trackbars of old and replaced with a new star control which we feel makes selecting values and ranges more immediately obvious at a fraction of the screen space. Black stars are negative values and Gold stars represent positive values.

The development team has also been optimizing the default adapter set. The ticker is now fully dockable (thanks Paul 😉 and behaves nicely compared to the floating topmost ticker before.

The ticker also has a number of modes, Multi-Item mode (which works just like a normal Ticker) and a fade mode, which was requested by the community as an additional optional transition to the Top and Right Sliding modes.

Optimizations to the cache and core importance algorithms have made Touchstone more responsive and more discerning about what items to interrupt the user with.

Age now more heavily effects the results of an item, and this has eliminated “item spam” the user got when they used to initially launch Touchstone. This is still a work-in-progress, but the effects are really effective.

We are preparing the build now, so it should be ready for release tonight or early tomorrow morning AEST (/me touches wood). Cross your fingers for us, and signup to get in on this awesome build.

“Almost beta” indeed 🙂

Ash Out