loader

What's my name again?

The OG convention

So, last year I wrote about naming conventions for Fabric items (What's in a name? Naming your Fabric artifacts — Advancing Analytics).

Just over a year on, and a lot has changed in Fabric world since then (not least that fact you’re now supposed to call them “items” and not “artifacts”).

Fabric is now Generally Available (GA) and there have also been new features introduced to make organising your workspaces easier. Somebody recently reached out to me and asked me if the blog from last year was still relevant. Is a naming convention still necessary? So now feels like a good time to write a follow up.

New feature number 1

We have coloured icons for our items! This really does help you identify items quicker at a glance.

coloured_icons

The level of thought Microsoft put into this is quite impressive! You can read the blog released alongside the feature here Fabric’s New Item Icon System | Microsoft Fabric Blog | Microsoft Fabric

New feature number 2

The toast of #FabCon was Folders. It seems likes it should have been a much more inconspicuous announcement, but it really resonated with the Fabric community, especially since it had been an outstanding feature request in the Power BI service since 2015. This is still a preview feature, but allows you to organise your items in a more hierarchical nature, even allowing nesting of folders down to 10 levels (that’s for information, not an invite).

How you decide to arrange those folders, is up to you. You could put common items in a folder, or perhaps at a higher granularity put common experiences in a folder (or nest item types inside experiences?) – a pattern I’m partial to is to group by process, so that you include ingestion items in one folder, processing items in another and presentation in one too.

folders_examples

As is often the case with preview features, it does have some rough edges (and how or if these will be resolved remains to be seen) – for instance, the folders are not supported for git integration. And personally I’ve had mixed results when deploying them via Fabric Deployment pipelines too.

You can find out more about folders here: Create folders in workspaces (preview) - Microsoft Fabric | Microsoft Learn

New feature number 3

The most recent new feature is Task Flows, which was another of the big announcements from #FabCon. These feel for me more primarily a workflow and lineage tool to help you build a solution, but the organising aspects are a nifty little side benefit.

Task Flows allow you to assign items to a task. These tasks give a nice colour coded reference in the List View of your workspace.

task_flow_tag

And you also get a  lineage-esque view of the dependencies between the tasks (though be warned, that lineage isn’t actually enforced, it’s more a visual aid)

taskflow

The Task Flows announcement blog is available here: Announcing the public preview of task flows in Microsoft Fabric | Microsoft Fabric Blog | Microsoft Fabric

 

Is a naming convention still needed?

So do all these new features make a naming convention superfluous?

Maybe I’m biased, but I think a naming convention is still relevant. The new icons make it easier to identify items, but grouping them together it still a good idea. So perhaps you could group them by folder then? Well yes, if you choose folder as the grouping choice for your items, then perhaps you can drop the type and/or experience parts of the recommended naming convention.

But if you do that, keeping some kind of index to represent the lineage of dependencies still makes some sense to me.

Or if you have used folders to group by process, then perhaps the index is less important, but then grouping by an experience or type prefix makes a nice idea. You could achieve this all by nesting and subdividing amongst several folders, but folder sprawl makes for a clunky navigation experience, so personally I’d advise against going more than 1 or 2 levels deep.

Retaining the full naming convention from the original blog perhaps gives you the most flexibility, allowing you to chop and change between different folder configurations whilst also making the intent of each item more explicit.

But as always with these things, there’s a huge dollop of “it depends” served on the side.

Feedback from the original blog was very positive, but I always intend advice around these kind of these things to be applied pragmatically as opposed to dogmatically. Personally I’m a stickler for consistency, so whether you choose the conventions I’ve suggested or not, some thought to a standardised approach to naming your Fabric items is definitely still something I’d recommend.

If you’ve still yet to dip your toe into the world of Fabric, Advancing Analytics would love to help give you an introduction. We run 3 day Fabric Hackathons that allow you to get your hands dirty with Fabric whilst exploring real world use cases for your data – you can find out more here: Download our Fabric Hackathon flyer (advancinganalytics.co.uk)

author profile

Author

Johnny Winter