Things we can do better than Windows

[Z98]

For someone familiar with I presume x86 assembly, besides learning C, I would suggest taking a look at the Intel or AMD manuals.

Intel: http://www.intel.com/products/processor/manuals/

The third link down, the combined systems programming one, would likely be the most helpful. And I would highly recommend reading the Windows Internals (latest edition, even if Alex gets money for it) book. It has a lot of information in it.

[nicamarvin2005]

Code: Select all

I am sure there are many things we can do that is better than how Windows does things

I'd be happy if Reactos was as Bad as Microsoft Windows...

[PurpleGurl]

Before this topic spans a few pages, allow me to tell you (and Z98) this: you basically agree with each other and there is little point discussing it further Maybe little clarification: If someone comes to irc "Hey, I have a problem, ReactOS BSODs when I <do something> on my 4-processor 32-cores-in-total machine.", but ROS devs say "OK, we cannot reproduce on any of our machines, the best we have is 3-processor 24-cores-in-total and it doesn't BSOD when we try to <do something>. It's not like we don't want to help you, but we can't as of yet without access to this kind of machine for debugging." That's the limit Z98 means.

I know. I understood all that several exchanges back precisely in the same way you explained it. We already said all we need to on this aspect of item #2, and I was already done with it (hence the reason I prefixed it with moving on - since I anticipated some would not).

As for the comment on doing things as "badly" as Windows, yes, that should be the goal for now. I am sure we will be even better in the long run, but we have to start somewhere. When we can run the majority of what Windows can run without serious trouble, we can continue. My point of this message thread is that there are a number of things we can do better than Windows and mistakes we don't have to make, all in ways that won't adversely affect compatibility.

Another brainstorm came to mind, though it won't really fit in with my list. What if the way some of Window's internal components talked to each other was inefficient or poorly implemented? That is a case where maybe we could have our own private calls to give better ways of doing those things, while still including the standard ones for software compatibility. However, I don't know that much about the internals to even know if this is applicable at all.

Z98, thank you for the link. I will have to seriously think about getting back into coding. It will be like starting from scratch for me. There are a number of things I consider show stoppers for me, and if I could help code them away, that would be great for everyone.

I've noticed that MJ Martin and Johannes Anderwald have been working very hard lately on the USB situation. That was one of the obstacles for me in using Reactos. For about 2 years, I was unable to run it on real hardware at all. I thought it would be different after the last version, but it wouldn't boot either. So I studied where it hung and figured it was the EHCI code. I first disabled all USB support, dug out a PS2 mouse, and Reactos booted fine. I noticed the mouse was very slow and that the mouse defaulted to slowest in the control panel. Then I rebooted, enabling just the USB 1.x support. The USB mouse was erratic and unusable, so I had to use the keyboard to power down. So I am very glad of the changes I see in the trunk and branches, at least 3 that may affect me. Now if more work was done on file systems and SMP support, I would be tickled pink (or would that be purple? ).

[bugboy]

Sorry for being off topic, but I just have to put a plug in here.
IRT Usb, please dont forget Johannes Anderwald. He has done so much of the work.
He deserves alot of credit for the progress being made.

mjmartin

[PurpleGurl]

Sorry for being off topic, but I just have to put a plug in here.
IRT Usb, please dont forget Johannes Anderwald. He has done so much of the work.
He deserves alot of credit for the progress being made.

mjmartin

Thank you. I knew I was leaving someone out and I edited my post to reflect his work on the USB situation too.

Now, I wonder if anyone else has things we can do better than Windows does them, and which most likely won't break compatibility?

[rosfanatic]

Probably asking the impossible, but: what if updates could be installed without requiring a reboot.

[vicmarcal]

Probably asking the impossible, but: what if updates could be installed without requiring a reboot.

Asking the impossible, yeah
There are some actions that needs a reboot to be performed. Substituting a critical part of the system needs a reboot to be performed(i.e a binary which is in use)

[PurpleGurl]

Probably asking the impossible, but: what if updates could be installed without requiring a reboot.

Asking the impossible, yeah
There are some actions that needs a reboot to be performed. Substituting a critical part of the system needs a reboot to be performed(i.e a binary which is in use)

Yes, but why not as minimal as possible or with the ability to resume the user environment in the exact state? By as minimal as possible, why not, for instance, reset just explorer.exe in a lot of cases? Windows installs a number of updates without a reboot, and some of the others that are not too deep could be done that way if Windows simply reloaded Explorer or logged off the user.

Speaking of update, it would be nice to have ROS to a point to where we can install it, and all the trunk revisions could be installed either automatically or on demand. Maybe the server could keep a checksum list and it could be checked against the system files installed and the ones not matching replaced. That could also help with a current version if there is a virus or corruption. And, for tweakers, the server could keep a checksum list of the previous official version and the current official version, and if there is no match to the current, previous, or trunk versions, then prompt the user on what to do since they could be custom files. At any rate, it would be nice to have an update system in place to where we can go ahead and install it and get new files as they become available without having to reinstall.

[boomhauer]

Yeah I've always wondered if updates require a reboot just because they weren't thought of in the initial design of the OS.... but, yes I fully understand how processes need to be restarted and files are locked etc, it does create a big web of dependencies which just gets to be impossible to try to manage. Evidence is how win7 seems to do better with needing reboots only when necessary, but even then still seems to require them quite frequently.

[sh4ring4n]

Probably asking the impossible, but: what if updates could be installed without requiring a reboot.

Asking the impossible, yeah
There are some actions that needs a reboot to be performed. Substituting a critical part of the system needs a reboot to be performed(i.e a binary which is in use)

Or it can leave the user alone until he shuts down on his own.

[jcalvarez]

Actually, forcing a reboot of the system is an easy way to make sure everything is properly initialised after a configuration change. Implementing a proper hot configuration update for some of the system components can be a real PITA.

Regards,

Jose