The future of the HQ Starter Kit?

[warrenbuckley]

Note

I initially raised this discussion in the MVP channel, but now moving here to get more 👀 on this and feedback to chat with the wider community and for HQ to chip in if needed.

---

Question for HQ engineering & or MGMT team

Is there any plans to rebuild the Starter Kit to use more modern approaches and try teach common practises of extending the back-office and building a site to go alongside the 17 LTS release?

There has been a ton of improvements in the CMS that do not showcase its power since it was originally built.

Does it make sense to kill the project off if its just going to get patched to work with the latest version or should it be considered to rework it to showcase Block List/Grid, Backoffice Block Previews and other features that make it a more useful project to help teach people and get up to speed with 17 and Bellissima.

Thoughts?
Curious to hear other people’s thoughts on the Starter Kit

[skttl]

I think "The Starter Kit" should be a demo. I think HQ should try not to have a definite best practice example (if they do, why are other options available?). In my opinion, best practice is just someones opinion. Eg. https://cultiv.nl/blog/tip-of-the-week-the-ultimate-site-structure-setup/, I like this way of structuring a site, but it's not objectively better than other structures.

[warrenbuckley]

Yeh I agree Soren, we all have different opinions and approaches and that's totally fine 🥰
Perhaps the Starter Kit, should be renamed to Demo or Example Site instead.

But I would imagine the engineering team must have some strong opinions and thoughts that perhaps others could learn from and get started a lot quicker from was my thinking.

The greatness of Umbraco is that we can do things a 101 ways, but as a beginner having so many choices and technical approaches can feel overwhelming and having some kind of starting point to refer would help people.

A simple example for context:
A block grid preview in Bellissima to display a picked image/media item.

I seem to have a varying different number of ways I could display an image from a GUID/key I have.

• Do I use an existing web-component in my component?
• Do I consume a context
• Do I call the MGMT API directly
• Something else ?

With perhaps an up to date starter kit or samples, I could quickly refer to it to see how the HQ engineering team might have solved the problem and for most people they would agree and follow the same process, as this is the same with the Lit, Vite & Typescript setup. As HQ uses this on the codebase internally it seems to be the preferred approach we as a community have tried to align with.

Again with the backoffice you don't have to use the same tech stack as Umbraco, but it makes sense to IMO.

For us people @skttl who have used Umbraco for a loooong time, we have our own preferences and opinions. But if I started over again today from fresh. I wouldn't know how or where to get started after installing and getting a blank/clean slate.

Curious to see what happens with this, as something I feel needs to be done.

[c9mb]

Yes, I also think it's fine for everyone to have different opinions, provide it's accepted that mine is the correct one...

[Luuk1983]

It think Soren is tackling a very important question: what is the purpose of the starter kit? I can think of a few things:

1. • A way to have some content to test some development stuff and packages. In that case it's simply a way to have some filling.
2. • A template for if you just need a quick, simple site. Much like how you can just spin up a simple WordPress website.
3. • A template with best practices to give you a head start with developing your own site. The idea is to build on this base.
4. • A demo, especially for new users, who are just curious what Umbraco looks and get some inspiration instead of just a blank site that can be frustrating if you don't know where to start.

Ofcourse some of these points are compatible with each other, I didn't mean you need to pick one of those four. However, they require different amounts of effort in my opinion.

For just having some content to test things, the current starter kit is fine. For a quick site, it's probably too limited. But for a demo to showcase the possibilities, I agree with Warren that the starter kit should have more body and showcase more examples. And it also takes more effort to keep it up to date. I don't feel the starter kit should be a base with best practices to build upon. In my experience, everybody has their own way of doing things and there is not a single 'best practice' I think.

[skttl]

Starter kits also have a limitation, at least in combination with tools like Deploy or uSync.

Typically a starter kit has a migration that installs all the content and schema. If you then install Deploy or uSync, and work on the site alone then that is fine.

Then you need to onboard a new developer, or you want a fresh clone/db of your site. The installer then (depending on your settings), import schema from Deploy/uSync, and then the migration installing the initial starterkit schema/content runs, which can conflict with what you have from Deploy/uSync.

For that reason we went away from having our own internal starterkit (with base schema, like a site node, content page etc), to simply have some automatic download of .uda files on project creation.

[skttl]

For packages, I tend to switch between The Starter Kit and Clean Starter Kit. I think they are both great options for this.

For the quick simple site, uSkinned has a lot of great options, or if it should be even simpler, then Clean Starter Kit works well (or umbBootstrap by Dean Leigh, haven't looked at that though.)

I have already mentioned my view on best practises - I don't think it's possible to make the "one ring to rule them all" for this.

For the demo usecase, the current Starter Kit is fine - although a bit dated. Could be a community effort to bring it up to 2026 level.

[warrenbuckley]

If the community was to own the project - we would still need someone at HQ to help steward/maintain the project if its to be released from HQ as an 'official' project.

I know I originally raised this internally in the MVP channel and @filipbech had some original thoughts & replies that they are thinking/considering what they do.

So I suppose until we hear more from HQ what their thoughts are, I am not sure there is much else to discuss for now. To ensure no one wastes any effort or precious free time until we know what the plans are.

[bjarnef]

I think the second video in the blog post here has a bit more content to demonstrate a real world use-case and also useful for testing in both PRs for CMS, packages or other extensions of backoffice:
https://umbraco.com/blog/release-candidate-umbraco-17/

It could possible demonstrate a bit more of the extensions points for collections (@nielslyngsoe demonstrated a bit of this at Codegarden), perhaps a dashboard and most basic features.

Another benefit is that it may be easier the catch any potential bugs.

[warrenbuckley]

Fingers crossed the video is then the Demo/Starter Kit replacement

[OwainJ]

I agree, I think the starter kit should defo be a demo, rather than a best practice guide, it should be a nice demo that shows off Umbraco and what it can do. IMO it's a good way of hooking in new devs/clients who've downloaded Umbraco for the first time to see what it's like.
Also, I think any/all starter kits listed on the marketplace, should be showcased on the install screen in some way, because otherwise, how does a newcomer know about them?