roadmap and priorities

Here you can discuss ReactOS related topics.

Moderator: Moderator Team

karlexceed
Posts: 454
Joined: Thu Jan 10, 2013 6:17 pm
Contact:

Re: roadmap and priorities

Post by karlexceed » Sat Dec 16, 2017 12:40 am

Are there more bugs than ever? Probably. There's also more code than ever. Regressions caused by re-writes are annoying but good in the long term. They expose flaws. Reducing duplicate and irrelevant bug reports is also a major issue.

The major problem facing ROS is, of course, development time/man-hours. There's simply not enough time being put into ROS to both continue making forward progress toward goals AND fix the new issues that are being reported daily. But every bug fixed is another step forward, just as important as a new feature.

I don't see this as a problem of "management". As has been said previously, the volunteers that work on ROS will work in their comfort zones. If a dev knows low-level hardware interfaces but not high-level graphics APIs, should we really insist that they focus their efforts on graphics? No, of course not.

So if there is a specific issue that needs to be fixed (like the memory manager...) and no one who commits code knows it well enough or has the time to invest, then that specific area should be where the money raised by ROS should go. Hire a dev to build out the MM. But then, the other issue... Where does that money come from?

It's a "chick and the egg" problem. People want to see their use cases supported by ROS in order to consider it seriously. If that happens, they might support ROS financially. But right now, they see alpha software that barely boots sometimes. Therefore, no money gets donated. Thus, no paid dev time gets added to ROS. The cycle continues.

What ROS needs is a patron. Or we need to do much much better at fundraising.

Or... Convince some more really awesome devs to contribute in their (hopefully ample) free time.

milon
Posts: 969
Joined: Sat Sep 05, 2009 9:26 pm

Re: roadmap and priorities

Post by milon » Tue Dec 19, 2017 4:32 pm

Michael Long wrote: But I partially agree on your point, that skilled people are necessary and you can't get them easily. I said "partially" because I still have hope that a good solution in the management could solve the problem but I don't think that it's .necessary to add more and more people (which we probably don't get anyway).
Either I'm confused or else you are. A "good solution in the management" (memory manager) is certainly critical to ReactOS's stability and viability, but it's certainly not going to drop out of the sky. The only way that good solution will appear is through the skilled people that we every much need. Or did I misunderstand you?
Michael Long wrote:Oh dear, I think this is difficult.
Yes. :)

Michael Long
Posts: 40
Joined: Wed Oct 25, 2017 7:51 pm

Re: roadmap and priorities

Post by Michael Long » Tue Dec 19, 2017 8:29 pm

I didn't mean the memory management when I said this. I ment the way the project is handled. But I am not 100% sure yet where exactly the error is. But I noticed the ugly symptoms like the growing amount of bugs, the feeling that more and more falls apart (like not being able to boot anymore, not being able to debug anymore, and so on) and the high interest among some developers in fixing issues which create a ton of new issues while neglecting more essential issues like those which have higher priorities in JIRA. If ReactOS would have more people then this process downwards might just be faster.

When I saw the problem the debian developers are facing (see the topic about the growing amount of bugs) I got scared that something like this could also kind of destroy the ReactOS project at a time where even more development time has been spent. The trend already exists. When I was trying to debug some issues I was facing such a massive amout of interfering bugs that I gave up (several times). I was not able to use my new monitor because ReactOS outputs at a not-supported resolution, so I had to use my old one. I was not able to use RS-232 to debug because ReactOS couldn't write to the RS-232 port. I was not able to use the log-file-creation-on-disc because of file corruption. I partially was not able to use the debug to screen option because the output on the screen was not correctly drawn. I was not able to use my keyboard and mouse because of an USB issue. I was not able to use the AHCI mode so I had to change BIOS settings in a more compatible way. There was just so much that no matter what I tried to debug the original issue I was always facing side issues which kept me from the original track.

I personally would prefer an operating system which can't do much yet, but that what it can, it can do reliably. I personally would also prefer an operating system which is simple yet, but has fully implemented the basic features of an operating system. But as I said, I can mention symptoms, I can warn of dangers, but I am still in the dark when it's about calling the root by it's name or even suggesting a solution.

dounet
Posts: 6
Joined: Sat Jan 17, 2009 4:47 pm

Re: roadmap and priorities

Post by dounet » Wed Dec 20, 2017 8:25 pm

Just one comment from me. We all want Reactos to progress and speed up. Let's put our money where our mouth is and let's donate, donate, and donate again.

Personally giving a monthly contribution and getting rewarded by seeing how the team progresses the OS with passion.

So if you have not donated recently - do donate. Simple.

learn_more
Developer
Posts: 143
Joined: Fri Dec 19, 2014 10:00 pm

Re: roadmap and priorities

Post by learn_more » Wed Dec 20, 2017 11:09 pm

Michael Long wrote:I didn't mean the memory management when I said this. I ment the way the project is handled. But I am not 100% sure yet where exactly the error is. But I noticed the ugly symptoms like the growing amount of bugs, the feeling that more and more falls apart (like not being able to boot anymore, not being able to debug anymore, and so on) and the high interest among some developers in fixing issues which create a ton of new issues while neglecting more essential issues like those which have higher priorities in JIRA. If ReactOS would have more people then this process downwards might just be faster.

When I saw the problem the debian developers are facing (see the topic about the growing amount of bugs) I got scared that something like this could also kind of destroy the ReactOS project at a time where even more development time has been spent. The trend already exists. When I was trying to debug some issues I was facing such a massive amout of interfering bugs that I gave up (several times). I was not able to use my new monitor because ReactOS outputs at a not-supported resolution, so I had to use my old one. I was not able to use RS-232 to debug because ReactOS couldn't write to the RS-232 port. I was not able to use the log-file-creation-on-disc because of file corruption. I partially was not able to use the debug to screen option because the output on the screen was not correctly drawn. I was not able to use my keyboard and mouse because of an USB issue. I was not able to use the AHCI mode so I had to change BIOS settings in a more compatible way. There was just so much that no matter what I tried to debug the original issue I was always facing side issues which kept me from the original track.

I personally would prefer an operating system which can't do much yet, but that what it can, it can do reliably. I personally would also prefer an operating system which is simple yet, but has fully implemented the basic features of an operating system. But as I said, I can mention symptoms, I can warn of dangers, but I am still in the dark when it's about calling the root by it's name or even suggesting a solution.
So I take it you volunteer to solve our issues by taking on the bugs that do not receive enough attention in your eyes?

ROCKNROLLKID
Posts: 293
Joined: Mon Oct 17, 2016 3:19 am
Contact:

Re: roadmap and priorities

Post by ROCKNROLLKID » Thu Dec 21, 2017 4:19 am

Well if you want me opinion, it doesn't really matter if they fix theme issues now or later, eventually it needs to get fixed. They could, lets say, focus all attention on getting hardware working, but if nothing works on it, then it is just as useless as it would be the other way around. Give it time as everything needs to eventually get fixed in order to have a stable operating system.

And as learn_more said, you can volunteer to submit code if you think you can take it on. Or else you could donate some money so they can hire a developer who can work on it.

zydon
Posts: 160
Joined: Tue Dec 18, 2007 9:03 am

Re: roadmap and priorities

Post by zydon » Wed Dec 27, 2017 7:55 am

I think development priorities in Open Source project were not even priority in it's daily operations. But, the priorities could be influenced by the progress of it's community activities and contributions. The Financial support, development resources, testing group and feedback are the thing we used to give as a community. Still the pace or the result of the support given, are the limited pace of the Developers willing to make.

Those are things that the ReactOS community cannot push it's pace or demand of it's progress. What the community can change is some kind of parallel activity progress which is could inspired the ReactOS developers not to missed out the opportunity to progress further when the community results provide the understanding of the communication between software and hardware and the solution to interface them up.

To simplify what it's all about, this forum need a new sub-forum for the community to do unofficial development experimental Lab or a dedicated hacking jobs that will spawn the result to understanding of the hardware and producing a driver for it or what ideas that giving a workaround on each ReactOS releases.

Some pointers about this sub-forum activities such as below:
  • The sub-forum could be called Unofficial TestBed or HackAway or whatever fit to criteria.
  • First priority could be discussion about how to develop a driver using ROSBE
  • Then Hardware resources used as references to communicate to the devices.
  • ROS Developers has no obligation to interact to the discussion but are welcome to guide to the right path whenever possible.
  • Allows members of the forum to upload picture files and if possible a compress formatted source codes and binary (with limited file size control) to be shared among members and become archive to the development progress. Link from outside will lost it's important resources through time.
  • Any crazy idea such as spawning ReactOS core to become the base of layer for future Android or whatever mobile OS could be good too to see how far members could materialize it with no rules attached.
It just an idea how community could contribute their BPU (Brain Processing Unit) output to change the condition of ReactOS progress and priorities.

karlexceed
Posts: 454
Joined: Thu Jan 10, 2013 6:17 pm
Contact:

Re: roadmap and priorities

Post by karlexceed » Wed Dec 27, 2017 4:32 pm

zydon wrote: this forum need a new sub-forum for the community to do unofficial development experimental Lab

[*] ROS Developers has no obligation to interact to the discussion but are welcome
[*] Allows members of the forum to upload ... source codes and binary
[*] Any crazy idea such as ... ReactOS core to become the base of layer for future Android or whatever mobile OS could be good too to see how far members could materialize it with no rules attached.
Thankfully, ROS is open source. You can create your own forum and fork ROS to do all of these things if you want to.

The trouble isn't that a non-dev can't get the source, but that many non-devs lack either the skills or motivation to do anything like that. If they could, I'd hope they would be a contributing dev for ROS.

middings
Posts: 1007
Joined: Tue May 07, 2013 9:18 pm
Location: California, USA

Re: roadmap and priorities

Post by middings » Wed Dec 27, 2017 11:18 pm

karlexceed wrote:The trouble isn't that a non-dev can't get the source, but that many non-devs lack either the skills or motivation to do anything like (unofficial development).
This is true. Non-developers (non-devs) like me must avoid getting carried away with big dreams for ReactOS that we ourselves can not implement, not even implement poorly. I try to remember Linus Torvalds's quip, "Talk is cheap. Show me the code." When I do remember that, I must then humbly shut my mouth because I have no code to give.

I try to do what little I can (mostly occasional testing on real hardware and making JIRA reports of bugs encountered), participate in the forums to answer questions or give advice about common problems encountered, and tweak the ReactOS Wiki a little bit (mostly to fix grammar). I would really like to contribute code but my programming skills are very rusty. I was never skilled in C and have no experience with operating system or thread-safe coding. I finally mastered compiling ReactOS from a script and Subversion (SVN). I appreciate the tutorials that others have written and the questions the knowledgeable people here have answered. The only patch I ever contributed was a spelling correction. For all else I try to follow the advice of Ludwig Wittgenstein, "Whereof we cannot speak, thereof we must remain silent."

Adcock
Posts: 231
Joined: Thu Jul 07, 2016 5:37 pm

Re: roadmap and priorities

Post by Adcock » Thu Dec 28, 2017 6:58 am

"Logic will get you from A to B. Imagination will take you everywhere."
Albert Einstein
That being said non devs should keep dreaming with open eyes beacuse :
Those who dream with their eyes open will make their dream come true.
D. H. Lawrence wrote something like that.
If someone wants to shut up then they are free to do so.
Why tell others to shut up ?
No one is pushing anyone.
This topic is about how development should take place.
But people are talking about unnecessary things.
How many times have the devs immediately implemented something that someone requested ?
That's almost never.
If yes then why are the innocent non devs being held responsible for incorrect or unexpected development.
My point is non devs might be talking too much but they aren't slowing the devs.
Devs do what they want.
So I think that if ReactOS is going in the wrong direction then non devs have nothing to do with that because they don't make the decisions.
Note : non devs = those who can't contribute to ReactOS development through computer programming.
I am a non dev too.

tverweij
Posts: 39
Joined: Wed Jun 18, 2008 7:17 pm
Contact:

Re: roadmap and priorities

Post by tverweij » Fri Dec 29, 2017 2:17 pm

What status of compatibility has ReactOS at this moment?
I mean that it reports 5.2 or Windows 2003 SP2, but it is absolutely not compatible with it at the current state.

The following versions of Windows NT existed:
NT 3.1
NT 3.5
NT 3.51
NT 4.0
Windows 2000 (NT 5.0)
Windows XP (NT 5.1)
Windows 2003 (NT 5.2)

Is ReactOs functional complete / compatible with one of these versions?
Is it known what has to be done to achieve functional compatibility with any of these versions?

If it is not even functional on par with NT3.1, that would then be a nice first step for a road map, with compatibility with the next version as the next step on the road map.
No time frames needed, but just an inventory what has to be done to get there. Just for the information of the devs, so they know what has to be done next to get any compatibility.

milon
Posts: 969
Joined: Sat Sep 05, 2009 9:26 pm

Re: roadmap and priorities

Post by milon » Fri Dec 29, 2017 3:19 pm

tverweij, ReactOS is an example of "flying the airplane while building it". It isn't fully compatible AFAIK with any NT architecture. The goal, as you pointed out, is full compatibility with NT 5.2, but we're not there yet. And it's not for a lack of roadmap either. Check JIRA - it's FULL of reports of what's missing/incomplete/etc. The devs know what needs finishing, and they're each contributing where they can. Creating a quality OS is a huge task. Creating an OS that's binary compatible with an existing (closed-source!) OS even more so. But you have to start somewhere, no?

EDIT - It's worth repeating that the devs aren't 40-hour salary code slaves. They're volunteers for the most part. They work on what they work on because they want to and because they can. Development wishlists are nice, but there's definitely no way to demand or enforce certain that things get priority.

tverweij
Posts: 39
Joined: Wed Jun 18, 2008 7:17 pm
Contact:

Re: roadmap and priorities

Post by tverweij » Fri Dec 29, 2017 3:48 pm

A quote from the wikipedia page of Windows NT 3.1 (https://en.wikipedia.org/wiki/Windows_NT_3.1):
250 programmers wrote 5.6 million lines of code; the development cost $150 million. In the last year of development, the team fixed more than 30,000 bugs.
So 250 full time developers took 3 years to build Windows NT 3.1 - that is 750 man-year - excluding testers.
And they started out with OS/2, a fully functional OS that they ported and extended to Windows NT.
If you take the development of OS/2 in account, the complete product took over 1500 man-year (excluding testers) to build.
And that was done with a proper design - first architects then development.

And the more complex a system becomes, the more (exponential) difficult it becomes to complete it.
Especially if there is no management, road map or design.

I am convinced that a project like Windows NT and ReactOS will never finish without a proper design and road map.
If you build the plane while flying, It will surely crash and never reach it's destination.


tverweij
Posts: 39
Joined: Wed Jun 18, 2008 7:17 pm
Contact:

Re: roadmap and priorities

Post by tverweij » Fri Dec 29, 2017 5:46 pm

Yes, I like that.

Why has nobody pointed out these pages when the first message in this thread came?

For reference (might be handy to have those links on the reactos homepage - project state):
1. Missing functionality: https://www.reactos.org/wiki/Missing_Re ... ctionality
2. Version status https://www.reactos.org/wiki/Version_Status
3. Next major version road map: https://www.reactos.org/wiki/0.5.0

For link 1: I think It would help if the project management would add priorities to this page - what points would give the most compatibly with Windows 2003 SP2.
For link 2: it looks like this page is not maintained since 2014 - it would help if it was updated

If I had to give the priorities, the highest would be (based on application compatibility and hardware support):
- WMI
- LSASS
- LDAP/ActiveDirectory
- Pinting (already started)
- Layered Windows (already started)
- Windows Driver framework
- Usermode Driver Framework

Post Reply

Who is online

Users browsing this forum: Ahrefs [Bot] and 2 guests