[openbeos] Re: Tracker icons
- From: Michael Phipps <mphipps1@xxxxxxxxxxxxxxxx>
- To: openbeos@xxxxxxxxxxxxx
- Date: Mon, 25 Jul 2005 07:43:32 -0400
On 2005-07-25 at 05:30:00 [-0400], Helmar Rudolph wrote:
> Axel Dörfler wrote:
>
> ... a sensible reply - almost as I would have expected it.
>
> > Please go into the details and define outdated. What part of an OS has
> > innovated in the past 10 years? Can you name a lot more than GUI?
>
> Yes, first and foremost it would be the GUI, control widgets for instance.
> But it's also stuff like USB, SCSI, wifi, access to advanced motherboard/chip
> features, the file system - things that prevent one from writing end user
> applications that then enhance the user experience, which in turn determines
> the usefulness of the system (OS+apps). Remember the original thread:
> Haiku/BeOS being slow with small files.
So... a couple of new widgets and a couple of new drivers. I guess we don't
have a whole lot of competition! :-) Seriously - Axel is right - not a whole
lot of really new and interesting stuff has been done.
> Fair enough. I also didn't say R2 will be done from scratch. I am saying that
> R1=R5 plus some improvements. R5 was released when? 2000? It's 2005 now, R1
> may be ready next year some time. Ok, 6 years instead of 7, so don't get too
> hung up on numbers. ;).
No, you didn't. But you seem to think that R5 is not enough. Personally, I
think that, feature wise, R5 is quite enough. I use it daily. Sure, more
drivers, better caching, faster kernel, more ram, better fonts would be nice...
Oh, wait - that is the "+" in R1=R5. :-)
> Fair enough, but do yourself the favour and don't become myopic here. We all
> know that apart from a few geeks, nobody gives a rat's arse about an
> operating system per se. All that matters is the end user experience, and
> that depends on the applications that are written for that operating system,
> their functionality, which in turn is linked to the GUI. But it's also the
> brains that went into the functionality in the first place, practical
> innovation that emerges as a result, and - like it or not - FOSS or not - the
> marketing of the OS and the applications written for it.
Of course this is true but I don't see where it has anything to do with the key
issue - start R2 now or when R1 is done.
> > Why not? When developers are making decisions it's usually a lot faster
> > than when other people are talking about options.
>
> That's exactly the problem with GE. Lots of people talking, nothing leading
> to timeous finality. OTOH, one thing where "developers making decisions" went
> horribly wrong on BeOS is the entire GUI, which is by far the worst and
> inconsistent among all other public options (known to me). No GUI consistency
> at all. Very little openess to "marketing" or "usability expert" input.
> Result: oblivion. Application oblivion == OS oblivion.
That's the craziest assertion that you have ever made! Least consistant GUI?
Hmmmmm... Windows has MDI and SDI; when you click the "X", you don't know if
your document or the whole app will close. Sometimes "new document" will create
a new base level window, sometimes it will be a window inside a window. No
warning, no clue to the user. Huh? I don't think that I even need to get into
any of the other OS's to show that this assertion is wrong.
> Again:
>
> Application oblivion == OS oblivion.
Think about how many app developers left when Be tanked. Developers can and
will live with a whole lot of warts. What they won't live with is
organizational instability - knowing that the OS that you are writing for will
still be available next week is a good thing.
> For sure, but you will soon see that to an extent you have to behave like a
> commercial entity, unless the outcome of your actions/effort is immaterial to
> you. If that is the case, cool. But then I suggest no one on this list uses
> the word 'success', because it still does have connotations.
Yes, to a degree. But because we are free from commercial requirements (like
turning a profit, for example), we can make longer term plans. Manpower
limitations, in fact, require us to do so. IF you have a plan that will help us
release R2 in the same timeframe that we could do R1 today, I would love to
here it.
> Success is eventually defined by the number of people using Haiku for their
> computing tasks. They will use application software, which will to a large
> extent be (at least semi-) commercial, be it $45 or $450. People want
> consistency; they want to know when new features arrive, when bugs are
> squashed, when the next version will enhance their computing experience even
> more.
People do want consistancy. I will take strong issue on the rest. When is
LongHorn, no, wait, Vista coming out? Date/time, please? Sometime next year? Or
mayber the year after? Historically, MS has had the WORST track record EVER in
planning and releasing versions. They, and pretty much they alone, are our
"competition".
> For that to even happen, the app developers rely on the OS they are writing
> their apps for. Even at the lowest level (FOSS written by volunteers), you
> can only build a momentum if there is a path of action, if there is a certain
> amount of reliability, consistency and professionalism. Unless, of course,
> you're happy spending all your time for maybe 5000 users worldwide.
The app developers need something to write for, period. R1 is that base that
they need. Once that is done, they develop apps, we develop R2 and blow
people's minds. This is why I have said so many times that I would love for R1
to stay quietly under the radar of most of the rest of the world. R1 is a huge
accomplishment. But I can't really justify to a lot of people that they should
switch. The apps aren't there. The ultimate cool new features aren't there.
That's what R2 is for.
> See the bigger picture, guys!
I think that I do!
> > Despite that we don't have the man power to do that, R2 builds on what we
> > have, it cannot really be worked on before R1 is ready.
>
> Why not? If R1==R5, then why can it not be worked on? Why can't the toolchain
> (which is partly at OS, partly at application level) approach be pursued? Why
> can't the other issues for discussion not be finalised, rather than being
> discussed ad nauseum but not leading anywhere? I don't expect miracles here.
> At the same time I think a different approach to GE and R2 would be
> beneficial.
For one simple reason. People hours. All of the developers are working on R1.
If more developers come along, they will work on R1, too. Is it fair to lock
Axeld in the basement and make him code R1 stuff and let Joe Noob develop stuff
for R2 (the fun stuff)?
> > I have plenty of ideas I want to put into one of our next releases after R1.
>
> Are _they_ being discussed on GE?
I am sure that some of them are. I am sure that there is stuff on Axeld's mind
that is not on there. Some of it may be a bit technical for GE (say, for
example, changing BFS from BTrees to SkipLists, just to make something up). 90%
of the people on GE wouldn't know what those are or how to evaluate the
difference. Which is OK - GE is about features, not implementation.
> > It's the nightmare of any good marketing and business people :-)
>
> Which, by and large, FOSS development is marginal at best. Which is why most
> FOSS developers are poor sods, who probably all dream about striking it rich
> (in their definition probably meaning "lots of money"), but who are clueless
> or helpless as how to actually effect it.
>
> FOSS and financial success are to a large extent mutually exclusive. FOSS can
> ONLY work if you build you own economic trading system around it. It cannot
> and will not work as long as you operate in the current capitalistic system.
Poor fool that Linus is, he seems to think that things are going pretty well
with FOSS. Others will build their own trading systems and value. That would be
a good part of the reason to use an MIT instead of GPL license. It encourages
just that.
> > But open source could not really work in another way; if the developers had
> > to do something they don't want to, why should they do it? For the money?
> > Well, there is none.
>
> But they all want it, right? Wouldn't they do more development if they got
> paid? Contradiction in terms, eh?
Sure. But I haven't seen any compelling evidence that doing any of the things
that you have suggested will get us to that point.
> > The only light at the end of the tunnel is that the developers might have a
> > common goal like "creating a great desktop OS" - that should make them
> > listen to the needs of others, and may make them adopt those.
>
> I must have written to half a dozen email client developers about an X-flag
> in the header that determines the reply status of an email. This setting will
> be picked up by the client and then displayed to the user. IOW, in the client
> you can check a box called "no reply necessary" when you write an email,
> indicating to the recipient that one doesn't expect a reply.
>
> Piss-easy to implement, very useful for those writing and receiving a lot of
> email. Did anyone listen? No. Maybe I have to patent it first for people to
> take note. ;) And that's just one of many.
But wouldn't EVERY mail client have to support it for it to be useful?
Inventing a standard is a very tough business.
> > But as you probably have experienced way more than just once, you'll have a
> > hard time with open source projects. It's hard to talk to us in these ways.
> > You must adapt and leave your box if you want to get through :-)
>
> I take that point. On the other hand I am a user first and foremost, and if
> you don't come up with stuff that woes me over from Windows (or whatever),
> then why should I bother? Now multiply me by a factor of 100.000.000+. You
> will be targeting people who already have computers, not newbies. You will
> have to make it dead-easy (1-click, anyone?) for them to convert all their
> existing data in order to be of any interest. You will have to provide
> features at OS and app level that make me say: "wow, I can't do that on
> [insert OS here], but I can do it on Haiku! And 'wow' how Haiku has taken
> over all my data from [insert OS here] without a glitch." And, and, and....
All of which take a whole lot of time. And should be discussed in detail on GE.
:-)
> In short: you want people to use Haiku and its applications for most of the
> time. This requires a modern OS, effective applications that are easy to
> learn and use (GUI consistency), and the communication of the benefits
> (rather than features) to the market at large
> - whether you are FOSS or not. Unless of course, you are happy with
> being/remaining marginal.
And there is a time for that. R1 is not it. R1 is about a base on which to
build. It is the foundation for the building, not the whole house!
> > R2 won't be the holy grail. The problem with wanting to change the way of
> > computing is that most of those that tried failed miserably; we'll be very
> > careful with that :-)
>
> Being careful is one thing, being immovable/paralysed another. Most tries at
> at application level anyway these days, but if the OS fails (ie on lots of
> small files or -as recently discussed on OSNEWS - with the inability to
> auto-read ID3 tags into the attributes), then not even the apps can save it.
I hardly think that we are paralyzed. I do agree that some of the "small
details" are critical. That is what BeOS has always been both brilliant and a
failure at.
> > We have limited capabilities, we
> > need to focus on our immediate goals. We let the forward thinker growing in
> > the background.
>
> I am not sure if 'growing' is the right word. 'Vegetating' or 'Neglecting' or
> 'Ignoring' may be more appropriate.
Yeah. Because none of the developers ever post on GE. I certainly never do.
Axel doesn't.
> > We actually do both, we're fulfilling our egoistical needs and desires by
> > providing us a more efficient computing experience.
>
> No, you are NOT providing a more efficient computing experience. You are
> copying something that in 2000 was still cool, but now no longer is. The
> practical innovation that needs to happen (on GE) WHILE you are doing an R5
> clone is just not taking place. Talk-shop. No structured path of action. And
> almost as a direct consequence there is no significant application
> development taking place, apart from TuneTracker, Refraction and Wonderbrush
> maybe.
So you think that if we told everyone "well, we won't have anything untill 2008
or 2009, but it will be WAY cooler than what you have now" we would have some
massive influx of new users?
> If you look at your universe as static, then yes, I understand that. But it
> isn't. At the end of the day it's really only about two things:
>
> 1) ideas/projects that get effected (finalised) and that lead to
> "success" (at whatever level or definition). [ Success can only happen if
> you have a reference point to compare it to, and this implies that you
> can separate the two, which means you have to finish stuff. ]
> 2) the effective/efficient communication of such ideas/projects
>
> Those two are at the base of everything. Everything.
>
> To end this thread, I am fully aware that this list here is populated by and
> large by developers who think and act in a particular way. Consequently any
> views from outside, especially if they are challenging the developer's
> current worldview, will be met with hostility - a not uncommon sign of
> underenlightenment, if I may throw some flamebait in here.
Now wait. This is the "you can't possibly understand me" arguement. Works
better from a teenager (which I happen to have one of). No, sorry, I won't let
her do it, nor will it fly here. What you must remember is that every developer
is also a user. They want it to be better, too. But they also realize that
releasing sooner is better than releasing better. To put it in marketing terms,
I think that you would have us chase an illusory market - the market of
developers who are unwilling to put in time building the basics and only will
work on cool new stuff.
> So... I didn't really expect any other response here. [Yours, Axel, was the
> most sensible, hence my reply.] At the end of the day, though, I sit here
> with enough money and time (and brainspace) on my hands to write all this.
> Given that the first two are usually in short supply with most followers of
> this list, I take that I may do something 'right' - for lack of a better
> word. As a result it seems stupid or foolish to me to discard my arguments
> out of hand. But it's your (pl.) choice, and thus your (pl.) reality.
Axel and I (and others) also have time, money and brain to respond. Does that,
IYHO, set us on equal ground?
> Success depends on a (w)holistic approach. If you think you've got it all
> sussed out yourself, well, you don't, otherwise you'd do things differently,
> and in this thread this is all about how GE is treated.
Oh. I see. We aren't HOLISTIC. Guys - THAT'S what the problem is. It isn't
time, it isn't bugs. It's a lack of seeing the whole picture. Now I get it.
<shakes his head>
> I'll leave you with this. What you make of it is entirely up to you. I just
> feel reminded of pretty much the same discussion I had on the comp.os.geos
> newsgroup in 1995/1996. You know where GEOS is now, don't you?
A *commercial* failure. Just like Be. Let me point out to you that Be's failure
was a lack of focus. 5 focus shifts in 5 years, or something similar to that.
What conclusions can you draw from that?
Other related posts:
