Multiwindow

[corvonero]

Being a MS fan (and, of course, an OSx critic), I strongly suggest you think the "multiwindow" feature request deeply, so this thread is to collect other opinions...

Multiple windows are a loss of time and effort (I repeat, this is my super-personal opinion) and probably useless in a Windows-oriented Desktop-like experience.

Under windows you see VERY VERY little applications that run with multiple windows.
Using multiple windows means having multiple menus, multiple "minimize to" and "maximize" options... that is something you really don't need under windows.

Chromeless windows (and transparent background windows) are a different story...

But I see multiple windows as a feature only directed to OSx style programs.

What you really use in windows is modal-windows (read: MDI and SDI - http://en.wikipedia.org/wiki/Multipl...ment_interface) which is something already brilliantly covered by sencha itself...

So, are people in this community really intrested in multiple windows?

True multiple window interface example: http://zone.ni.com/devzone/jsp/large...ng&language=it

MDI - SDI example: http://qt-project.org/doc/qt-4.8/mainwindows-mdi.html

[jarrednicholls]

I agree, this feature definitely serves OS X more than it does Windows or Linux. In fact, it's down right required on OS X to support the idea of "multi-instance", where on Windows or Linux running multiple processes does the job.

I'll be the first to admit that I'm not the grand master of use cases But I'm sure there is some use case where having multiple windows that are completely segregated would allow for an easier UX when multiple displays/monitors are involved. The first that comes to mind is an application like Adobe Photoshop. If the docked panes weren't undockable and floatable windows, I think working with Photoshop would become incredibly hard, especially if you had more than one monitor and wanted to move the panes off to another display so your canvas area is maximized. Maybe this use case is a stretch, but still valid when speaking generally about dockable/floatable panes. I bet one day Architect would take advantage of this feature.

I do agree that MDI (tabbed and/or modal windows inside parent container) and dockable interfaces are very well covered by Ext JS, and they do cover the majority of use cases. Architect uses these to its advantage very well (especially in v2.2). What Architect can't do on OS X though, is allow for the user to open multiple projects at the same time. Multi-windows feature will allow for this.

We're very interested to see what this poll reveals! In our private beta testing group, we did get a lot of requests for multi-window management.

[elgs]

I personally don't think the multiwindow is worth the complexities it brings in. Instead I would vote to put client filesystem read(I saw it's done from the examples)/write to higher priority.

[Lukappa]

I think this poll is funny... what are we talking about?
In my opinion a great tool should meet almost all developers needs.

@corvonero
About windows...

Being a MS fan (and, of course, an OSx critic), I strongly suggest you think the "multiwindow" feature request deeply, so this thread is to collect other opinions...

You can't build a software depending on your own opinions and feelings (excluding you are doing it only for you).

Multiple windows are a loss of time and effort (I repeat, this is my super-personal opinion) and probably useless in a Windows-oriented Desktop-like experience.

The same as before.
About effort/loss of time is not true.
2 external native windows use less resources than 2 native windows wrapped into another native window (so 3 native windows) - with "native" I mean not "Ext.window.Window"
I don't want to say that one way is worse than the other way, it depends on what you need.

What you really use in windows is modal-windows (read: MDI and SDI - http://en.wikipedia.org/wiki/Multipl...ment_interface) which is something already brilliantly covered by sencha itself...

How can you do if you have dual monitor and want to move 1 window in 1 monitor and the other window into the other monitor? Simply you can't with Ext.window

Anyway I think you are right if you are talking about web development.... but for this we have browsers yet.

If we are talking about native applications, the Sencha packager, in my opinion, should meet every possibile way to handle windows / component / files / notifications and every possibile interface to the OS (without prejudice on brands).

[jarrednicholls]

If we are talking about native applications, the Sencha packager, in my opinion, should meet every possibile way to handle windows / component / files / notifications and every possibile interface to the OS (without prejudice on brands).

This is what we aim to do in time. The idea of these polls is generally good in terms of helping to prioritize the roadmap. I don't want to discourage anyone from doing this; there's really no harm.

However let it be known: multi-window management IS on the list, and it IS being implemented for our next major release. The decision's already been made and being carried out. We're also taking care of some of the issues and requests we've already heard here on the forums, including better binary file i/o. I don't expect this poll to reverse any decisions, but it does perk up my curiosity and creating conversation.

[corvonero]

This is what we aim to do in time. The idea of these polls is generally good in terms of helping to prioritize the roadmap. I don't want to discourage anyone from doing this; there's really no harm.

However let it be known: multi-window management IS on the list, and it IS being implemented for our next major release. The decision's already been made and being carried out. We're also taking care of some of the issues and requests we've already heard here on the forums, including better binary file i/o. I don't expect this poll to reverse any decisions, but it does perk up my curiosity and creating conversation.

Thanks for clarifying what I took for obvious: the poll was meant for checking what people consider more important.

I personally think PDF support would be MUCH more important than Multiwindow.

And, of course, I didn't even think for a split second that the poll would make sencha change its mind... I just thought that after reading about 3 threads for Multiwindow, one could think it over in favour for more important things...

(( and I also agree that ALL is better than SOME... but as it's clear to me, we're looking at a early stage program... and there's still a long way to go for a "ALL-DONE" release ))

M.

[corvonero]

I also think (correct me if I am wrong...) that sencha is not trying to create something that will replace java, c#, .net, cocoa... this is just a great, fast and easy way to "fill the gap" between html and OS development.

But if someone is planning to make a sencha (or html5) based Photoshop replacement I think he should REALLY, REALLY think the project over...

So, I personally think that a good milestone would be a packager release that emulates 100% what a good browser does so that we don't need to make 2 different projects or, even, implement workarounds all over our projects.

I see desktop packager as a great opportunity to build simple to medium applications fast. Not something too CPU consuming or disk intensive... (again, correct me if I am wrong). I would not opt for JS to build an SQL engine!

I also think the packager is a GREAT way to protect JS code... really, a GREAT way!

So, honestly, things like PDF need greater attention...

[jarrednicholls]

Agreed that the Desktop Packager should run your project 1:1 as it would in Google Chromium. Given that the Desktop Packager *is* Chromium, we're 98% of the way there. With a little feature detection, it's nice to have the extra integration into the host system for complementary functionality above what HTML5 currently offers. 2 different projects or workarounds shouldn't (and aren't) necessary. I also agree that some features are arguably more important than others. We appreciate input on these matters. We have our next version planned out and are already into implementation; I was just noting that.

Though I understand your point...clearly, building a SQL engine in Desktop Packager would absolutely be silly and that's not a use case we or the web is trying to solve. But don't underestimate the web! Building a (subset of) photoshop application, for example, in web technologies is absolutely possible and plausible. Canvas, WebGL, hardware accelerated computing, etc. I've myself created such an application over 3 years ago when canvas was brand new. The tech is so much better now-a-days. New standards are coming that will bring OpenCL (WebCL) to the web for paralleled computing, too. We also have some fun features coming soon specifically to Desktop Packager that will aid users in this area. I wouldn't set your sights so low...

So, while Desktop Packager isn't aiming to replace Java, .NET, Cocoa etc., it is most definitely aiming to be an option for web developers to take advantage of the web platform intimately along with tight integrations into the host platforms, and develop the large majority of the use cases and desktop applications that one could create using all native tech. If there was ever a validation of this goal, see Microsoft's investment in Windows 8 and its 100% JS API coverage and support for writing first-class windows applications in 100% web tech.

Again, a PDF viewer is a great feature request! I've posted in these forums a nice workaround to load PDFs in the user's default browser using the Ion.util.openUrl method, in the meantime.

[Lukappa]

I perfectly agree with you corvonero... of course we were talking just about multi window... not about complexity of app.
So, Sencha team, go on with this packager!