WebSite Information > Article Discussions
[Arca Noae] So, what would you like to see next?
Martin Iturbide:
I just posted this news: [Arca Noae] So, what would you like to see next?
Before sending anything to Arca Noae, I will like to discuss something first.
When you are asked "what would you like to see next?" on this platform everyone (included me) will reply first "I want to have the driver for "my device" working. And like I discussed it (long discussions that I don't want to repeat) it is only short term. In four years we will be asking for the same thing for newer hardware.
What I will like to see is a "try", an "intention", an "attempt", a "spark" to have a long term future for the platform (which some people thinks we don't have a future, and I also don't want to repeat that long discussion).
I will really like to see small, but continued effort to clone OS/2 Presentation Manager since it is an important component in the OS/2-eCS experience. OS/2 PM API is big, but the idea is to start small selecting which are the components that needs update/replacement first.
- Select little by little small goals of the API to clone (Device Functions, Window functions? or by DLL?)
- Trying to clone PM to run over eCS-OS/2 Warp 4.5x first (we don't want Linux kernel on the way yet)
- Some tricks to mask PM and use Qt on the background may be interesting (but I don't know at which level Qt requires PM).
- Make it a open source project, so it can "adds up" to the future of the platform and do not became future close source abandonware.
- Consolidate and use the code of other open source projects that can help like FreePM, OSFree, PM++ lib.
- Replace eCS-OS2 PM DLLs one by one, little by little.
And by the way, I'm not saying "Arca Noae, just go on and do it while I watch", I want to help on this with whatever is on my hands.
Any other ideas of what can it be a "long term project" for the platform?
Jean-Yves:
I'm not sure whether this is doable, but I'd love to see an updated/improved VirtualBox OS/2 Guest Additions package that does not require manual steps and which allows the use of Panorama or SNAP (not sure about licensing there)
I've given up trying to get eCS running on real hardware and actually rather prefer it in a VM
Olafur Gunnlaugsson:
--- Quote from: Martin Iturbide on November 25, 2014, 03:48:03 pm ---I will really like to see small, but continued effort to clone OS/2 Presentation Manager since it is an important component in the OS/2-eCS experience. OS/2 PM API is big, but the idea is to start small selecting which are the components that needs update/replacement first.
--- End quote ---
Clone the WPS, the PM is 80's code and concepts that needs to be retired. The effort of more or less straight cloning of OS/2 like OSFree is going to have the same end result as Haiku had in the BeOS world, by the time it becomes a workable system, everyone has lost interest and moved on, all vendors have disappeared and the community not big enough to support itself.
Something like this:
http://research.microsoft.com/en-us/projects/barrelfish/
With a x86 emulation layer or virtualiser if on an x86 system, on top of the OS that allows us to run eCS-OS/2 as the front end of the system while we code a modern WPS replacement and tool set, if the base OS becomes outdated or unsupported or in some way undesirable, simply port the emulation layer and WPS replacement/Modern OS/2 API to a new system.
Martin Iturbide:
Hi Olafur.
I will also support anybody that wants to create a open source clone of WPS, I will also jump into that ship too.
First I will like to remember two things:
- PM is the OS/2 GUI
- WPS is the desktop shell. The objects that organize the GUI to create the desktop interpretation (based on a real's world desktop) of the OS. WPS requires SOM and PM to show itself as it is.
But why, if I had to choose a priority, I will choose PM first (warning, this is the interpretation I do based in my knowledge) :
1) I tried to list all the applications that uses WPS - here it is the list on the wiki. The WPS applications that are available are minimal compared to the PM applications that we have available on this platform. Maybe XWorkplace is the most evolved WPS based project with more classes, but it does not compare to the quantity of applications that requires PM.
2) PM is the real GUI of the OS/2 platform and that will be a definitive real good step on open sourcing the platform and have a continue future for it.
Other alternative like trying to mask the PM methods and behind it use Qt will be interesting too. For example if you replace the method that calls the "Save As.." window from PM and call the one from Qt it can be an interesting trick if it can be done, so the developer do not have to write all the method, only write the "conversion mask" and reuse the Qt code (It requires knowledge of both things Qt and PM). But the trick will have to be that the PM applications think they are using PM while at the end is Qt, and that the application does not stops working/breaks.
Other road is not to clone PM and to try to port other GUI over OS/2-eCS. Unity, Razor-Qt or even the not-yet-created Quantum OS (that uses Qt 5 and QML). On that road what will have is two different and incompatible GUIs and there will be a lot of effort to try start creating applications for that GUI until the majority of users migrate (or start to like) that new GUI. It will create a lot of friction (just like the initial eCS 1.0 Vs OS/2 Warp 4.5 friction that we used to have on 2000), but maybe in the long term we can have an individual OS/2 users with the new GUI and the legacy users (corporate) that requires PM because they have a PM application that they refuse to leave and does not have an equivalent on the new GUI . So it can be a interesting bumpy road :) ....and maybe someone will say.. "this is the OS/2-eCS community, isn't PM an important part of the platform?"
My main concern is that we may have ideas, but we don't have any developers, money or corporate support that are willing to jump into any of this possibilities. I only wish I can do anything to try to convince someone with more resource (money/knowledge/time) like Arca Noae, Bitwise works, Mensys/XEU, or individual developer groups to do something for the long term (and open) future of the platform. They are doing a great job on what the do, but nobody wants (or do not have the time) to start touching the things under the OS/2 hood. The excuse I get is "if it is not broken, don't fix it"...but to me it means "boiling the frog" (the metaphor)
Mike La Martina:
As a long time OS/2 user and developer, I am curious about the perceived need to replace PM. Exactly which benefits would follow from this?
i have the same question about WPS?
I think that we could benefit from some advanced native controls, but beyond that why should we replace that which has worked so well for so long?
Navigation
[0] Message Index
[#] Next page
Go to full version