Remove .Net Standard

[Mimetis]

It's becoming harder to maintain a .NET Standard version.

Are you still using .NET Standard ?
Is it ok to remove .NET Standard supports and supports only .NET versions > .NET 6.0 ?

[VagueGit]

UWP devs are stuck on .NET Standard. .NET 9 support for UWP is currently in preview

Maybe support .NET Standard for 12 months after .NET 9 support for UWP is released? This gives UWP devs time to upgrade.

[Mimetis]

oh nice, didn't see that .NET 9 is going to support UWP !

[VagueGit]

The idea is to make it easier for devs to migrate from UWP to Win UI and MAUI, so presumably there will be at time when UWP goes out of support. So we'll be looking for DMS support for Win UI and MAUI please as those tools become available.

[NancyFolsom]

I am stuck with standard for the foreseeable future. Especially since my client is a U.S. government agency and who knows how funding will be available for a redesign that drops the framework I'm tied to.

[JScrivo]

I am stuck on framework 4.7.2 for my client app foreseeably for the next 2-3 years so it could create compatibility issues for me. :(

[NancyFolsom]

Me, too. As much as I'd like to move them off 4.7, they have to find budget for it. I've been working on them for a few years now to plan and work on an upgrade, but it'll be a complete rewrite which is hard for customers to stomach. Esp. hard at the moment in the U.S.

Edited to add: I mentioned DMS likely dropping support for classic .NET for necessity, which is 100% reasonable, IMO. That might help.

[Mimetis]

@NancyFolsom If it can help your customer move forward, that's a good reason :D

[NancyFolsom]

They're not opposed to upgrading just because they don't want to deal with it. If there's no budget to rewrite the entire application, then there is simply not the budget, and it would take time.

[jantoineqci]

Maybe you can start Dotmim.Sync 2.0 with no .net standard support and have a chance to rework everything the way you want without worrying about backwards compatibility (but have a migration guide). Short-term you would have some complains but long term you would have a better library.

[Mimetis]

I don't have any plans to migrate to a new whole new architecture ;)

[VagueGit]

Short-term you would have some complains

fails to grasp the damage done to critical services by breaking backwards compatibility

[jantoineqci]

fails to grasp the damage done to critical services by breaking backwards compatibility

fails to think long term in which demoralized by the daunting task of ever growing backwards compatibility complexities. The maintainers of this open-source project may give up, leaving the project un-maintained and un-patched for said critical services.

Please keep this an objective discussion.

[VagueGit]

My apologies @jantoineqci . No offense intended. Your choice of words appeared to trivialize the impact of breaking backward compatibility. I have misunderstood and thank you for calling me out.

Now I understand you are aware there are significant apps around the world using DMS in government and in industry where breaking backward compatibility will bring more harm than just a few complaints. That is why, although apps do become more challenging to maintain as they mature, complete rewrites of significant apps, like DMS, are rare.

You will also be aware that even though DMS is a large and mature app, the maintainers show no signs of flagging in their support. We need only look to the significant development that has taken place over the last year. This is possible because maintainers follow recommended practice 'add unit tests and refactor'

That's not to say we wouldn't welcome you @jantoineqci or others taking on the task of DMS 2.0, as long as DMS 1.x continued to be maintained.

EDIT 1: I should have included a reference to this article on Things You Should Never Do by Joel Spolsky (formerly CEO Stack Overflow). It's a hugely entertaining and highly regarded article often referred to as 'Death By Rewrite' It's a golden oldie

[Mimetis]

Now I understand you are aware there are significant apps around the world using DMS in government and in industry

Really ? :)

the maintainers show no signs of flagging in their support

The maintainer.

DMS 2.0

What feature(s) can bring us to such a big refactoring and a new version that is so different that it could be called DMS 2 ? (and that is not back compatible with actual DMS )

The only thing I think of, would be probably a full peer to peer sync framework, where there is no server / client typography and where each instance can be either server or client.
to be honest, I've already think of this kind of new architecture, but I failed so far to find the correct algorithm