Look this article is very important for .:Haiku:. proyect.
After reading that editorial, one can assume that Eugenia is a really really direct person. :lol: 
I agree on most except for her gloomy conclusion for Haiku.
I don’t see any problem with haiku other than the following:
No apparent milestones laid out, where everyone works towards the next point. Perhaps a poor example, is that I get the impression all the easy stuff has been done. For example, BButton is 100% complete, yet the underlying GUI messaging framework is not completed yet.
Just not enough detail / status updates. Job lists etc. I’d really love to help out doing some small jobs (to start), but it’s impossible to know where to start, what the task is dependant on, what it hooks into, etc.
No guide to installing the source on BeOS, building, making, and a walkthrough of say for example testing an object / part / kit. EG, how did the BButton code get tested? If things are like this are publicised, lots of people will see that they can do that, and you could recruit many more people who may not have the skill and creativity to invent, but have the skill to utilise, apply, and scrutinise, and improve / bug fix.
No apparent direction in the kernel and other big teams. Really all tasks should be more detailed, and have a decent chunk of text to describe the required work. Not a two liner that says “implement x and y”. That just doesn’t motivate 99% percent of the population. Only that small minority can picture what the task involves.
Tasks marked are un-assigned and assigned. Two issues: What tasks are un-assigned because they are dependant on other tasks? What assigned tasks are stagnated, and are in need of re-assignment, re-commitment, or a kick up the arse?
Not enough emphasis on target and results. Personally I feel that all efforts should be focused on the kernel, storage and networking. Effectively I would want to see the kernel boot up, and allow you to telnet in, run commands and execute command line applications. This is the basics for the remainder of the OS. Fair enough I see that people working on the appserver are “appserver” people, networking people are net people. But there is a lot of developer / team inter-mingling and perhaps time and efforts going off in a tangent.
When you look on the website, the team pages do not link the actual active team pages. For example the appserver, interface kit, and kernel all have separate mailing lists, websites, and private hidey-holes. There should be decisive direction from the project leaders to either make the Haiku forums the main repository of communication, or the website be kept tightly bound to the real goings on wherever they may reside. The kernel forum has only 1 post, and it was someone asking how they submit a patch! Someone was finding it hard to contribute!
The website team must recognise that they are much more than a web host, they must participate with the teams, have close communication with team leaders to pass on as much detailed information as possible to the public. In fact the communication aspect of the website, should involve just as much work as the OS itself.
I don’t think the Haiku project will stall or fail. I just want to see a more focused approach, gain more assistance, and communicate better. I would want to be in a position where I was of use, but at the moment I’m frustrated by a lack of not knowing what or how to do.
I’m just a hacker. Not a psychic, or mad-scientist like most of the haiku team. When I’m at work, I write flow charts, boring software specifications, lots of notes, scribbles, and sub-tasks for even a small job . Some links to the relevant part of the bebook, and a few googled sources for theory, and practical considerations goes a long way to motivating people, and giving them the believe that they could achieve completing a task. I’m sure a lot of this has already been done by the teams at some point during the beginnings.
Sheesh … go re-read your post dude and then do some searching on the website, the mailing list archives and the forums.
You complain about stuff that has been answered time and time again.
As for Eugenia … she’s the FUD monster !
Sikosis wrote:
Sheesh ... go re-read your post dude and then do some searching on the website, the mailing list archives and the forums.You complain about stuff that has been answered time and time again.
As for Eugenia … she’s the FUD monster !
So? It’s not easily findable for the majority. And way to go on the attutude. You must make a great team leader. :roll:
>No apparent milestones laid out
http://haiku-os.org/learn.php?mode=goals
http://haiku-os.org/learn.php?mode=status
>Just not enough detail / status updates
Are you on the mailing list ? Ever logged onto the #irc channel and see the CIA bot mentioning the latest cvs commits.
http://cia.navi.cx/stats/project/OpenBeOS
>No guide to installing the source on BeOS
http://haiku-os.org/learn.php?mode=nsl_view&id=52
Questions should be directed to the mailing list if regarding kit’s test modules.
>No apparent direction in the kernel and other big teams.
Micheal Phipps has talked about this before that we need more developers in the low level stuff. At present, we have Axeld working on the kernel when time permits and DW is busy on the app_server along with 1 or 2 others.
What direction would you like ? We’ve discussed in great length about how we can get an idea of where these components are at, to convey to the public.
>Tasks marked are un-assigned and assigned
That’s up to the individual team leaders to look after their own team assignments, etc. Speaking as the Pref Lead, we had over 200 people signed up as developers and got code from about 5-10 of them. I only actively hear from about 2-3 of them.
>Not enough emphasis on target and results. Personally I feel that all efforts should be focused on the kernel, storage and networking.
Sorry to sound like Monkey boy but … developers, developers, developer s …
>When you look on the website, the team pages do not link the actual active team pages.
True, there are a couple of sites where information is kept at present (DW’s for one). This is just until the admins have time to look at the system that the web team have put in place.
>The website team
Yes Kurtis attends our regular weekly meetings and also is on our admin only mailing list - but traditionally news articles have been written up by the team lead of said story.
>I don’t think the Haiku project will stall or fail.
Nope neither do I. That’s why Eugenia loves to say it will. She wants to prove herself right. I guess we’ll just have to prove her wrong.
>I’m just a hacker. Not a psychic, or mad-scientist like most of the haiku team.
Sorry if I was a little abupt, but I just feel like all I do these days is answer the same questions rather than contributing to the project.
http://haiku-os.org/learn.php?mode=goals
http://haiku-os.org/learn.php?mode=status
Not really pertinent to helping a developer. I meant detailed milestones for teams, not the OS, or an entire kit being alpha, beta… I mean stuff like boot kernel -> list of tasks, load drivers -> list of tasks (PCI, ISA, I2C, IDE, USB stack), load kernel TCP/IP…
Are you on the mailing list ? Ever logged onto the #irc channel and see the CIA bot mentioning the latest cvs commits.
http://cia.navi.cx/stats/project/OpenBeOS
I’m on one of the mailing lists know of two others, and maybe 4 dead others. Also know about the Cia site. But where are these on this site?
>No guide to installing the source on BeOSI know about this, but it hides in a newsletter, why is it not in the tutorials or resources section?
Questions should be directed to the mailing list if regarding kit's test modules.Wouldn’t it be good if each team page had a active new section of it’s own, where for example, you could see current activity, what works really, in plain view. IE “in appserver server land today”, you can download and build (using .proj), run app X, and it will connect to the test server, and shows the appserver serving up a few windows, with proper z ordering (screenshot attached)… It could be as simple as linking directly to a sticky in the forums. Instead of digging through a mailing list.
>No apparent direction in the kernel and other big teams.
Micheal Phipps has talked about this before that we need more developers in the low level stuff. At present, we have Axeld working on the kernel when time permits and DW is busy on the app_server along with 1 or 2 others.
What direction would you like ? We’ve discussed in great length about how we can get an idea of where these components are at, to convey to the public.
Simple plain view communication if milestone. Big bold text for example in the kernel team page
Current Milestone-> bootloader.
Next Milestone-> load kernel drivers from HD
Followed by list of needed developer skills. Checking up on the way milestones are currently about 30 or so small task slowly climb up a percent or two now and again. You don’t get a good visual peception of the progress that is being made. IE, for appserver / interface the individial tasks could all be moved into separtate milestones, for basic framework, messaging / threading, controls.
That’s up to the individual team leaders to look after their own team assignments, etc. Speaking as the Pref Lead, we had over 200 people signed up as developers and got code from about 5-10 of them. I only actively hear from about 2-3 of them.
yes that’s not good, although to be expected to a small degree. Perhaps not that bad though. My pet hate is agreeing to sign up not knowing what tasks you will be expected todo. I wouldn’t agree to a task unless I had the sort of info like on DW’s site: http://www.beemulated.net/network/darkwyrm/openbeos/serverdocs/toc.htm
The majority of internet coders aren’t really self-motivted / self-starters. Myself included, most need a good kick, and some basic framework laidout. There’s nothing worse than a blank piece of paper. I think this is real reason why there is a very concentrated developer base, because the information is locked up in the heads of the genius’ and perhaps doesn’t flow down to good.
Sorry if I was a little abupt, but I just feel like all I do these days is answer the same questions rather than contributing to the project.
PR, is always necessary but evil work.
To euan:
You have found the words to describe what I’ve been feeling.
Sikosis wrote:
There's an irc channel? First I've heard of it.
Are you on the mailing list ? Ever logged onto the #irc channel and see the CIA bot mentioning the latest cvs commits.
And the mailing list link is in Experience->Resources, which is hardly where a developer will look for it. Perhaps moving it to Develop->Resources.
euan wrote:
Not really pertinent to helping a developer. I meant detailed milestones for teams, not the OS, or an entire kit being alpha, beta... I mean stuff like boot kernel -> list of tasks, load drivers -> list of tasks (PCI, ISA, I2C, IDE, USB stack), load kernel TCP/IP...
This is actually the job of the team leads, but no one wants to take the time to do it usually. Keep reading.
euan wrote:
I'm on one of the mailing lists know of two others, and maybe 4 dead others. Also know about the Cia site. But where are these on this site?
This is actually my fault. During the entire year long development session I spent on the site, I never realized, and no one ever commented that a: the irc channel isn’t really listed anywhere noticable, and b: there’s no contact page. I’m actually in the process of working on this stuff, and it’s not like its hard, I’ve just been busy with work.
euan wrote:
I know about this, but it hides in a newsletter, why is it not in the tutorials or resources section?
There’s an outstanding warrant for articles exactly like this, but no one has taken it upon themselves to write one, even when I specifically asked for things just like this.
euan wrote:
Wouldn't it be good if each team page had a active new section of it's own, where for example, you could see current activity, what works really, in plain view.
Another thing i’m working on is devoting webspace to individual team sites… kind of like how darkwyrm has the app_server status site on his own. But again, this takes time because the team leads don’t normally have the time to do things like this. This would be a perfect thing to volunteer for, as well as coding.
euan wrote:
I hate PR work =)
PR, is always necessary but evil work. ;)
And as far as the alleged “editorial” … I kind of promised I wouldn’t say too much about it… at least on the osnews site anyway…
pulls right hand away from keyboard with left hand
euan wrote:
Sikosis wrote:Sheesh ... go re-read your post dude and then do some searching on the website, the mailing list archives and the forums.You complain about stuff that has been answered time and time again.
As for Eugenia … she’s the FUD monster !
So? It’s not easily findable for the majority. And way to go on the attutude. You must make a great team leader. :roll:
Way to go on the insult, too. Sikosis is often in the forums answering newbie questions (many of which are repeated, as many in your post were). The preferences team is probably the worst for people turnover, and must take a hell of a lot of organising.
I for one think you’re doing a great job Sikosis.
Doh! Knew I meant to say something about the editorial too!
BeDoper has a great take on it:
http://www.bedoper.com/bedoper/2004/twentyninth.htm
Unfortunately ELQ is also of the opinion R1 is a clone of R5, again I call for Haiku to advertise the differences more. Still, I know you won’t suddenly set a whole collection of “we have to catch up with MacOS X in 15 months” type of goals - the way you’re going with a fixed end point for R1 (R5 compatibility) brings the most fore-planning to any open source project ever IMHO, and is absolutely the correct way to go.
Compare interest and progress in Haiku with those projects with vaguely defined goals - “it’s going to be like BeOS, but well, just better, and it should run linux apps too, etc, etc”. I hope the other projects prove me wrong and come out with something, but having such a well understood goal really helps Haiku devs to focus on the tasks needed to get R1 out of the door. And Eug might reckon it’s too late, but until she can point to an easy-to-use, professional, consisent, open source OS I’ll have to respectfully disagree.
Keep it up guys.
And oh yeh, happy birthday 
Simon
well, I’ve finally done it. This post has broken my silence.
I hope this doesn’t sound like a cop out but here I go.
Both sides have a point. WalterCon has opened a lot of sleeping users eyes with new hope and excitment due to many amazing anouncments (Java and name change, etc…) This has brought forth a wave of excited and eager “novice” coders, developers and people who just wanna help. All this new blood has also brought to the surface a need for a new philosophy. No longer is it going to be possable to tell people to just start. The experiance level of these newcomers (not all mind you) is not to the point that they can browse some driver code and hop on the band wagon. There is no longer room in the code to just blindly start hacking. It has gotten to a point that it needs focus. New coders need direction to the holes that others have passed over on their way to their stomping grounds.
On the other side of it. Newcomers do need to take some time and read through whats there. With some simple prompting, ie a page on the site listing much of what sikosis listed, could look into getting started. There are not a lot of devs that have time to code let alone hand hold new coders through their first steps.
I believe these problems are just growing pains of a maturing project. A lot of good will come of it if we handle it correctly. Snapping at people and continually saying “its on the site” and “just start” isn’t going to win many new allies. If all these questions have been answered so many times before, shouldn’t the repeated asking of them make you wonder if perhaps a new delivery of the answers is in order. Asking newbies to dig through mailing lists, old articals, and irc just to get answers to “what can i do” or “what is needed” isn’t a very welcoming proposition. There needs to be a new developers area to the site that contains (in no specific order):
- tutorials on replacing old BeOS with new Haiku kits, controls, and apps
- information on how to test Haiku kits, controls and apps
- directions on where to ask for help and what help already exists.
- <enter you idea here>
Perhaps another way of doing this would be a wiki. Then people could post when they can, new developers that have gotten started can post for newer new comers. I would recommend a bit more structured and formal approch to it then the bebits one. I get lost in there all the time 
One last note. This new wave is (for the most part) less developer and more “home comer”/ex-user. What would really benifit them and us (dev’s) IMHO is a dedacated section on bebits. I think that if we did that and people could truly see the amount of usable code in the tree it would bring a better sense of how far we have come. I personally still have trouble determinning what is usable from most of the tree.