Poll - Removing Desk

41

Bad question (IMHO)

If it generally takes 10 to 12 man hours each month to support a given desk UI and there are now 3 different UI available, then that means there is now potentially a 30 to 36 man hour investment every month REQUIRED just to keep pace.

That is a ridiculous requirement to place on an all volunteer team. If they essentially loose a whole week on man hours to maintain the UI then suddenly the ability to move the rest of the project forward is gravely impacted.

If I were on the UI team, I would come up with the simplest to maintain method of displaying the contents and set a standard for it. Then if anyone wants to make their own mouse trap, they can in knowing that the basic tenants of the UI groups work will always look and feel the same.

This would allow essentially a version proof way for others to invest time into developing any fancy mouse trap of an interface they may want and know it is not likely to break on the next version change of the core system.

Don’t get me wrong here… I absolutely LOVE the old icon desk!!

But I would much rather have volunteer developer time optimized so that I can pay for the interface that I want with a paid developer and know that it will likely continue towork on every subsequent version upgrade.

Just my opinion (and you know what they say about opinions)

BKM

42

“Perhaps” is not an accurate assessment of the community responsibility.

This IS where the community responsibility rests.

However, it is EQUALLY the responsibility of Frappe to create the simplest way possible to display the contents and keep it standardized so that others can fill in the parts that take up a lot of time.

If that is an alphabetical ‘tagged’ list of contents then that is what it should always be and never changed so that others can write long term code to support whatever user interface they want and THEN we can see where the real community efforts will pay off with wonderful and beautiful interfaces that are yet to be imagined.

Frappe should start treating this like more of a universal open source project. Back away from the stuff that causes arguments like UI and just set the table for others to fill in that part on a more permanent basis.

Believe me, if this were done, there would be 3 or more brand new interfaces in a very short time and all of this nonsense bickering would be over already.

It might also stop some of this “how do I know if my contributions will be accepted” nonsense as well. you will know if your contribution is accepted if it can stand on it’s own as a new UI. The best will stand the longest and be the most used. The ones that don’t have the best community interest will likely never be used. So instead of complaining about your UI pull request, you would be able to see your own results based on how well your interface is adopted.

None of this can happen though unless Frappe sets the table for it to happen.

BKM

1 Like
43

I had not previously seen this thread. Very interesting stuff.

My first thought: there is already the REST API.
https://frappe.io/docs/user/en/guides/integration/rest_api

So to a certain extent, what BK is suggesting already exists. If I wanted to, I could create my own UI right now. Make my screens look however I want. And for every Create-Read-Update-Delete (CRUD) operation, just call the ERPNext (Frappe) API.

Now, I’m trivializing the effort. In my opinion, building a user interface is actually quite tricky. I’ve spent tons of time building a small, corporate site from scratch. Even with frameworks like Vue.js, Bulma, and Bootstrap? It’s hard work. If it wasn’t, everyone would be a front-end web designer.

Also, we know that some ERPNext business logic is residing in the current UI/front-end JavaScript code. Instead of in the back-end. That would need to be handled.

But it could be done. Today.

@bkm, correct me if I’m wrong. But I “think” what you’re suggesting is something in-between?

  • The individual DocType screens (Customer, Vendor, Sales Order) would be universal and the same.
  • But the means of accessing those screens (Desk) would be different? Different launchpads?

I feel that too can be done, right now. Spin up a web page. Make it look however you want. Start adding buttons, lists, boxes. Each would points to a different ERPNext page. Too simplistic?

Or are you thinking ERPNext should be like Debian/Ubuntu? Where the core functionality is the same. But you can download completely different desktop environments (Gnome, Unity, Mate, Cinnamon, XFCE)? That would certainly be awesome. But take some serious effort. Anyone who’s done this knows it can be buggy. I’ve broken my Ubuntu a couple of times, installing so-called “compatible” desktop environments.

Final thought. I’m struggling to think of any other ERP system that let’s you completely re-skin the UI and desktop environment. So we cannot say ERPNext isn’t following a universal standard for ERPs.

Again, interesting thread. But I’m not convinced it’s Frappe’s responsibility to make ERPNext modular like this. If someone made such a thing? Then I think it’s Frappe’s responsibility, as maintainers, to incorporate the contribution into ERPNext. But that’s all.

1 Like
44

That is essentially correct. Since they already have ‘links’ in the sidebar of any screen beyond the purple desk, those can be listed and assigned tags that never change. The tags would either always continue to point to the same doctype/function or be retired completely.

Then anyone (possibly even using a simple WYSIWYG web developer program) could try their hand at building their own ‘desk’ by simply calling the tags from their simplistic UI in order to get to any function as if it were from the old/new desk. They could even make their own goofy icons if they wanted.

And like you, this is trivializing things a bit, but you get my drift here.

While true, my pointing to Frappe to ‘set the table’ for this was to get them out of the the middle of the argument once and for all. I figured it would be in their best interest to set this up and let the UI titans that want to argue their cases fight it out in the arena of “who attracts the mot users” to their UI.

Rather than giving the developers a platform to complain about competing user interests, they can do this one thing and get themselves out of the problem they profess to hate so much.

BKM

1 Like
45

Any tool this scale has to make some hard decisions.

I totally agree with this. The reason Linux never made desktop is that there was no org that was willing to “own” the desktop. Maybe the likes of HP, IBM, Dell were arm twisted by M$ into not pushing an OSS desktop.

Our (maintainers’) aspiration for ERPNext is that we intend to “own” the complete user experience and that will mean taking opinionated decisions. Also we have indicated multiple times that “maintainers” is not restricted to Frappe. Anyone who becomes a reliable contributor overtime and is aligned with the vision of ERPNext is welcome.

If our decisions are wrong, we will face consequences (like forks) and if our decisions are right, everyone will win. Those who claim to be “helpers” should really think of aligning with themselves with this vision, because it is what it is.

4 Likes
46

My company is still a long way from actually deploying ERPNext anywhere, but in our prototyping we’re really interested in this approach. So far it seems promising. In a company of several hundred people, only very few would ever see standard forms or list views. Virtually all interaction would be done with Pages.