Mendeley Desktop: The About Dialogue (and the Refactor)


New About Dialogue and it’s even closable on Mac! How on earth did we get to this point?

Let me tell you a story…

Since my last post the Mendeley Desktop team has been very busy indeed!

Our “rewrite of some of the internal Mendeley code” has turned into a rewrite of almost all the code.

Why are we rewriting our code you may ask? Like many other small companies, Mendeley started with their software being written by external contractors. By the time I started in May 2008 a significant amount of code had been written which was then ported to Linux and Mac. What started life as a prototype had turned into a product which was then released to the world when we hit open beta. This means we lacked a solid architecture, any real documentation, coding standards or unit testing.

In the past few months the team has rewritten basically everything except the metadata extraction and the Citation Style Language parser (although these two have both been improved also and will probably be incrementally rewritten for further releases). As a result, we now have a much smaller, easier to read, unit-tested, documented, faster and just plain better codebase by any software engineering metric you would care to throw at it. My slightly obsessive insistence that the team meets the coding standards document and our continuous integration tool has resulted in a far higher quality product.

You might be asking why on earth you should care about everything I said above. Well the answers are in the features/bugfixes that you’ll see in the new release:

  • Better performance and lower memory usage
  • Adding sub-groups
  • Folder monitoring
  • Encrypted data transfer
  • Only uses standard HTTP ports (i.e. 80 and 443) and uses your system proxy settings
  • Less interface slowdown on network/import operations
  • A closable “About” window on Mac, as featured at the top of this post (No, I’m not joking. To close it in 0.5.9 or below, press Escape)
  • More native and more usable user interface
  • More traditional Mac packaging

Anyway, I hope that is enough to get you excited about the next release. We will be retaining feature parity with 0.5.9 (i.e. no features currently existing in 0.5.9. will be dropped in 0.6.0) and 0.6.0 will fix a lot of outstanding bugs with 0.5.9.

Mendeley Desktop 0.6.0: Coming soon!

Mendeley Desktop: The MVC strikes back

Firstly, thanks to all of you who have filed feature requests or flagged up bugs on our bug tracker, your feedback has been really useful.

After thinking about how to best accommodate your needs we’ve deciding to do a rewrite of some of the internal Mendeley code in order to get it running snappier and work better with larger libraries.

As a result of some work Fred has done on his music player we’ve decided that using Qt’s MVC framework maps extremely well to our needs and should bring us far greater performance when dealing with large libraries as well as keeping the code cleaner, better separated and making it much easier to write unit tests with decent coverage levels.

So where are we with this at the moment? Currently we have implemented most of the new non-GUI code with just our local/remote database code to finish. For those interested in this sort of thing, we are 100% documented with Doxygen (both private and public members) and have every class’s method unit-tested with around 85% test coverage.

Without revealing too much you should be excited about this and some of the new features that will see the light of day around the same time as the new, better-performing internal code. This will hopefully mean we can feasible support much larger collections than currently with a slimmer/faster application and quicker bug turnaround with less regressions than our previous releases.

Get excited, I am! 😀

Worst. Result. Ever. Brilliant!

By chance, I stumbled across One Big Lab yesterday, a very interesting blog on Open Science maintained by (as far as I can tell) four Stanford bioinformatics Ph.D. students. One of the many gems to be discovered there is a series of t-shirt designs called “Worst. Result. Ever.”:

You’ve been there, done that. Spent hours, days, weeks… months?… just to discover that your hypothesis (or “hope-othesis”) is completely wrong. Finished a data analysis only to see that what you’ve just produced can only be described as the Worst. Result. Ever. […]

Each one is named after the hapless student who had the pleasure of seeing something very much like it in their own research.

I’ve had nightmares of the Magda, and once pulled a Bernie, too. Once the shirts become available, buying them will support the PSB workshop on Open Science in… what? Hawaii?! I need to go there!

This week's roundup

Busy week again at Mendeley HQ, so just a few short notes of what we’ve been up to. Besides Mike joining us – and of course ongoing development work on both Mendeley Desktop and Mendeley Web – we’re still on the lookout for a new office.

We did see a few interesting ones, such as an old abandoned jewellers’ workshop in a house full of gold and diamond traders close to Farringdon. The place was oozing character (and big safes hidden behind the shelves) while being relatively affordable, but it was in dire need of refurbishment which most likely couldn’t have been completed in time for our moving.

Unrelatedly, I wanted to show you an interesting bit pertaining to the design of Mendeley. Yesterday, I set out to create the new default profile picture shown on a researcher’s web profile before an own photo is uploaded. This is how it went down:

You may notice that I’m certainly not an artist and that I couldn’t draw a straight line if my life depended on it. However, a certain co-founder of mine is even worse. That’s why Paul and I do most of the design work on Mendeley.

As a researcher, you’ve certainly heard the phrase that research is “10% inspiration and 90% perspiration”. I believe that this ratio might be skewed even more heavily towards perspiration when you’re designing things – because it will almost never look good on the first attempt. Case in point: Getting to a decent-looking profile picture didn’t just take the five attempts above, but at least another 20 unsightly failures.

For the 10% inspiration part, I’d recommend heading over to Smashing Magazine – they’re simply amazing.