It all started with a Spark

Behold, the unbelievable power of statistics!

Hello! Hello, is there anyone there? Yes, I have to admit I completely forgot all about this place. Indeed it was an impromptu discussion around blogging with a fellow dev whilst en route to Bristol on Thursday night that actually reminded me I even still have this blog. As the reason for the journey South was for something techy though, it seems as good an excuse as any to blow off the dust here for a change, but I can’t promise it’ll be a regular thing.

Friday 6th March saw the second year of Umbraco Spark at M-Shed in Bristol, billed as the innovation conference for Umbraco devs that want to find out what’s going on in the world of Umbraco. I’ve long since intended to get to some of the bigger Umbraco events having only made it to local meetups so far. But dates have a habit of clashing for me, so seeing this come up on the calendar was an opportunity not to be missed.

Having needlessly started the day worried about being late (I do this often), what better way to address this than by taking a look at the very concept of time itself! Jon Skeet started things off with a broader C# talk about the limits of DateTime, and ways of mitigating risks with dates and times. I’m sure I can’t have been alone in not knowing just how many different calendar standards there actually are, but it’s definitely left things that’ll swirl around in the back of my head if I ever have to do anything which is heavily time dependent.

After that first talk, things got a lot more Umbraco-ey and split off into multiple tracks. What with always being keen to have ecommerce ideas up my sleeve, having looked at Tea Commerce in the past, and having seen an awful lot of teasing Twitter posts from Matt Brailsford over recent weeks, I had to drop in on his talk about the new Vendr ecommerce platform. What was shown on the day certainly looked impressive, and importantly actually still felt like it was part of the CMS as opposed to an external system tacked on. It’s not quite ready for full release yet, but it’s very close and definitely looks like it’ll be a good system, and importantly works with Umbraco 8! So one to keep on my radar.

After a short session of discussion and Q&A for a live umbraCoffee (I make a very brief cameo in the crowd shot at 35 seconds), the next talk came courtesy of Paul Seal and Tim Payne. Taking a look at performance of querying data in v8, this one really proved to be an eye opener. Many of the things I know around querying for nodes goes off best practice advice and knowledge I’ve been using since late Umbraco 4 days, and I have to admit to not having really delved into how much it has changed in v8. Traditionally on very large nodesets, the logic was to use descendants and children querying quite sparingly, rely a lot more on examine instead, and then cache the hell out of things. Whilst some of this still applies, for example if making more complicated text searches, the raw statistics from testing proved that with the newly improved caching layer in Umbraco 8, avoiding descendants or adding on your own application cache can actually be slower than using what’s available natively. This is one of those things I’m going to be taking away, doing some of my own testing on in the coming days, and then considering actually refactoring some code. A definite plus of an event like this is that it really does get you thinking in ways you might completely miss otherwise.

After a lunch break spent talking with fellow Umbracians whilst overlooking the sunny waters of Bristol, it was back in to the afternoon track. Bjarke Berg gave us a runthrough of the work that’s in progress so far to bring Umbraco over to .net core. It’s certainly clear how big a project this is, but it’s good to see a very firm strategy for it. All too often I’ve seen people just bandy about terms (bonus points always available if you can get ‘microservice architecture’ mentioned) without really much of a plan beyond that, and that definitely didn’t seem to be the case here.

Following that, I switched track to catch Emma Garland‘s talk on tying together Umbraco and LUIS. A lot of my own time lately has been spent working with LUIS and Microsoft Bot Framework, and tying both of them into data from Umbraco, so this was a topic right up my street. Despite the amount of time I’ve spent in some of these systems recently, it’s still made me aware of a few extra features I may be able to make use of… as well as confirming I’m not the only one to suffer a long running internal debate on whether we should pronounce it Loo-is or Loo-ey!

The last of the track talks of the day I caught was from Marcin Zajkowski on performance testing Umbraco. Performance is an often overlooked metric, with all too many going down the ‘well it works fine on my dev machine when I’m the only one accessing it’, despite the huge business difference a well performant site actually makes. After going through a lot of the business stats on this, Marcin ran through lots of various tools, websites and tips on ways to benchmark and check performance for sites. Some I already knew. Others I’d heard of but not used. But it gave a good spread, and for both frontend and backend. Indeed there were so many that he had to get through in what was quite a tight timeslot I struggled a little to keep up with some of them, so may have to dig out any slides that appear later for this one.

The last talk of the day was from Niels Hartvig himself as he brought the ‘Not a Keynote’ speech. This one took a little more of a wider look at innovation and exploration, and the dangers involved (as no one ever rewards the failures). Looking at how this related to Umbraco when it launched over a decade ago, how challenges have changed since, and what the future may hold for software in general. As is often the case with keynotes (that-are-not-keynotes), the whole talk took an angle designed to pose more questions and get people thinking rather than provide any easy answers. Although it did conclude the day with an encouragement of how we should try collectively to democratise digital for all – a good message!

And with that, it was all over, seemingly almost as quickly as it had begun, although in reality 8 hours had passed. After such a long day, I was sadly too exhausted to manage the Friday night of socialising too. Although from what I’d seen of people in the community I barely knew just during the day, I don’t doubt this would have been a good time too. It’s very often said that Umbraco is the Friendly Community, and this whole event is the first larger event where I’ve been able to see this first hand. I’ve already left feeling more inspired to try and get more involved with the community myself, and will definitely try and get back again for any event in 2021.

One thought on “It all started with a Spark

Comments are closed.