[Rant] AOS4 on AOne slow? Perhaps, perhaps not. | ANN.lu |
Posted on 23-Sep-2003 09:53 GMT by Fabio Alemagna | 105 comments View flat View list |
Ok, I haven't had a chance to look at those videos until a few minutes ago, and I thought I'd give some reasons for which it appears slow, basign my judgments on what I see and what I know about the AmigaOS internals.
First of all, I must say that, although there was perhaps nothing more to see, those videos only show one little part of the GUI system, which, from what we know, could perhaps be the only one a bit flawed.
In essence, what we are shown are 2 things: the speed with which Reaction GUIs are drawn (in Solid-Resize and GUI-Reactivity), and the way (a faulty one - but read on) opaque window moving is implemented (in Opaque-Move).
Indeed, Reaction GUIs are drawn quite slowly, but I can't help but notice that Reaction GUIs have always been inherently slow, at least on my UAE setup, much slower than any other GUI, even MUI. The reason for this can be researched in the fact that Reaction uses a completely different approach than MUI, which is also the reason for which I greatly dislike Reaction in favour of MUI. However, it's obvious that the redraw is quite slow also considering Reaction's faults, and this is possibly due to the fact that the gfx library was emulated, along with a basic chipset support (some blitter thingies, like the BLTDONE flag in DMACONR register), which surely slowed everything down a lot.
Using this argument to say that AOS4 is at the same stage MOS was 2 years ago is pure flamebait for mainly 2 reasons: 2 years ago MOS already had a native gfx subsystem (which renders the comparision useless, although it might seem that it makes the situation even worse for AOS4), and also because this gfx subsystem is a temporary one, and we don't know at which state of development the new one is (but I reckon it's close to completion, or on the way to it). In any case, such comparisions are meaningless because AOS4 has now things which MOS didn't have 2 years ago, and vice versa.
About the opaque moving, instead, it can be noticed that there's a flaw in its implementation, which basically makes it *very hard* to *impossible* for the damaged windows to refresh themselves until the movement has stopped. Read Georg Steger's explanation of this phenomenon here .
|
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 51 of 105 | ANN.lu |
Posted by hooligan/dcs on 23-Sep-2003 12:47 GMT | In reply to Comment 48 (Anonymous): 6. MUI was chosen to be the MOS GUI, therefore it must suck? :-) |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 52 of 105 | ANN.lu |
Posted by Anonymous on 23-Sep-2003 12:50 GMT | In reply to Comment 49 (Kjetil): > Hypersion should add counters inn the function to measure the number of time function calls are used (only for debugging)
You are the man ;-) Counters in function calls. You have heard of a profiler, haven't you? |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 53 of 105 | ANN.lu |
Posted by Fabio Alemagna on 23-Sep-2003 12:51 GMT | In reply to Comment 38 (Anonymous): > If you learned to differentiate between opinions and facts in how you present
> your arguments people might feel like being more constructive towards you Fabio.
Shut up, wanadoo. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 54 of 105 | ANN.lu |
Posted by Fabio Alemagna on 23-Sep-2003 12:51 GMT | In reply to Comment 39 (catohagen): > no, i doubt filling/flooding this site with random 'thoughts' by random normal
> people
Then stop it. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 55 of 105 | ANN.lu |
Posted by samface on 23-Sep-2003 13:04 GMT | In reply to Comment 42 (krize): This is a *perfect* example of how some people simply don't read what I write. How could I possibly make the example I did in comment #15 if I did not use MUI? Of course I have MUI since there are (unfortunately) so many applications that requires it. I stated that it is slow in comparison to Reaction, not that I hate it to the point of refusing to use any application that requires it. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 56 of 105 | ANN.lu |
Posted by samface on 23-Sep-2003 13:14 GMT | In reply to Comment 47 (Eva): Ever considered the wild and crazy idea that I would have BOTH a PC and an Amiga? |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 57 of 105 | ANN.lu |
Posted by itix on 23-Sep-2003 13:16 GMT | In reply to Comment 48 (Anonymous): > MUI is not a standard. It does not follow the AmigaOS Classes system (official way of extending the Amiga's GUI), Wrong. a) original C= BOOPSI classes were too limitedb) MUI uses standard BOOPSI interface ClassAct (->Reaction) did extend original Intuition classes though. > It's SLOW. Gadtools, in comparison, is about 2-3 times faster in screen redraws alone. But gadtools interface is *very* difficult to do especially if you support different fonts (others than Topaz/8). GadTools is fast but how many gadtools based web browsers do exist? > It installs a whole new set of library (and more) files on your hard drive--do Windows .DLL files come to mind? LIBS: directory on my AmigaOS partition contains 293 files alone. > MUI uses it's own system of screen modes, which means needless hassle, more wasted time. You can run applications on WB screen if you wish. Programmer can bypass PSI as well if needed. > Gadtools responds/redraws faster, doesn't require extra support files, and still looks good. Also text mode apps in MS-DOS look good. Arent gadtools gadgets in OS4 BOOPSI gadgets now? |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 58 of 105 | ANN.lu |
Posted by Don Cox on 23-Sep-2003 13:44 GMT | In reply to Comment 48 (Anonymous): "Consistency in a GUI is important; any professional software designer will tell you this."
It is a common saying, but IMO consistency is overrated. Programs differ so greatly in their purposes that it makes more sense to give each the best GUI for the job, especially when they run on separate screens.
For example, Lightwave (on Amiga) and TVPaint have completely different interfaces. Both work well, both suit the type of program, and you can run both programs at once and flip between them without (IMO) getting confused.
I don't think it matters at all if some programs are MUI and others are Gadtools or Reaction. Nor does it matter if they have different colour schemes, or different shaped arrows on the scroll bars, or different fonts on screen - any more than it matters if two books on a shelf have different typefaces and layouts. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 59 of 105 | ANN.lu |
Posted by MarkTime on 23-Sep-2003 13:54 GMT | you know OS4 slowness seems to be the thread of the day, but while I've
enjoyed the valid criticism of team Hyperion over the past few years, because
I think constructive criticism is beneficial....I don't happen to believe
this slowness is anything important.
It's just a snapshot in the development cycle...it doesn't have to be slow
by the time its released.
Of course, I always have to say, I'm still disappointed that this isn't completed and on sale since many months ago...
but that aside, debug code, emulated code, just poorly written code, its all makes things slow, but that doesn't mean it won't be fixed before release....of course it will be fixed, what point would there be to releasing an unusable product...they have to have a sellable product, and I'm sure they will....the question is when. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 60 of 105 | ANN.lu |
Posted by Fabio Alemagna on 23-Sep-2003 13:57 GMT | In reply to Comment 59 (MarkTime): > you know OS4 slowness seems to be the thread of the day, but while I've
> enjoyed the valid criticism of team Hyperion over the past few years, because
> I think constructive criticism is beneficial....I don't happen to believe
> this slowness is anything important.
Well, that's exactly the point I made. This post of mine was a pro-AmigaOS post and I even got attacked by the same old trolls (not talking of you). Look what can brainwashing do... |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 61 of 105 | ANN.lu |
Posted by MIKE on 23-Sep-2003 14:10 GMT | In reply to Comment 14 (Rassilon): Well, no, the world wouldn't of produced the Model T, 3 years after the Mustang hit the market. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 62 of 105 | ANN.lu |
Posted by BrianK on 23-Sep-2003 14:11 GMT | In reply to Comment 58 (Don Cox): It is a common saying, but IMO consistency is overrated. Programs differ so greatly in their purposes that it makes more sense to give each the best GUI for the job, especially when they run on separate screens.
-----
Don't know what someone can say to that. I think GUI consistency isn't overrated. There's nothing worse then hunting for a Save or Print key because someone doesn't want to use a standard as they feel their GUI is better and will save time. It may do that but it increases the learning curve for the end user.
I'm going to compliment us all and assume the average Amiga User is more techie/DIY then the average PC user. However, when you give a system to Grandma and give her 5 different programs with 5 different interfaces she'd rather learn something else then this 'complicated' machine. However, if you can be consistent so once she learns how to Save in one program it's exactly the same in each other she's more likely to use the system and try new things.
While each program does have it's own purpose they all have common factors such as printing, saving, opening and closing files, etc. The way you get to these should be consistent on a platform.
Guess I just care for an organized standardized system instead of a willy-nilly system w/o consistency in it's applications. Might as well just build a base system w/ only a text interface and let each application handle it's own Graphic interface and integration. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 63 of 105 | ANN.lu |
Posted by brotheris on 23-Sep-2003 14:17 GMT | In reply to Comment 50 (Crumb // AAT): he was speaking about aos39-2, OS4 is not public. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 64 of 105 | ANN.lu |
Posted by Some Farker on 23-Sep-2003 15:13 GMT | In reply to Comment 58 (Don Cox): "It is a common saying, but IMO consistency is overrated. Programs differ so greatly in their purposes that it makes more sense to give each the best GUI for the job, especially when they run on separate screens.
For example, Lightwave (on Amiga) and TVPaint have completely different interfaces. Both work well, both suit the type of program, and you can run both programs at once and flip between them without (IMO) getting confused."
IYO. Every other OS has consistent UI design. The two most successful ones do, as well. Tell me: why *shouldn't* ctrl-v paste in every application? Why *shouldn't* alt activate the pulldown menus?
"I don't think it matters at all if some programs are MUI and others are Gadtools or Reaction. Nor does it matter if they have different colour schemes, or different shaped arrows on the scroll bars, or different fonts on screen - any more than it matters if two books on a shelf have different typefaces and layouts."
There used to be a UI Hall of Shame on the web; I wish I could find the URL and post it. You just might change your mind if you saw it. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 65 of 105 | ANN.lu |
Posted by BrianK on 23-Sep-2003 15:45 GMT | In reply to Comment 64 (Some Farker): http://www.umlchina.com/GUI/ShameRecent.htm
http://pixelcentric.net/x-shame/
Are a couple of good sites for bad GUI stuff.
If you're a programmer this should be one of your light required readings -- GUI for Dummies...
http://www.amazon.com/exec/obidos/ASIN/0764502131/isysinformationa/102-1487241-1940136 |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 66 of 105 | ANN.lu |
Posted by Treke on 23-Sep-2003 15:55 GMT | In reply to Comment 52 (Anonymous): Hi
Profiling ... to some degreee ...
At some stages we (at our company) using the parallel port. Benefits:
1. It is fast. The value is almost immediately there (1-2 microsecond)
2. You see real latency values of your application, including task switching.
3. Its' easy to track the values with osciloscopes.
4. Gives you more exact measurement values (just try to measure an operation in microseconds on a PC, where the timer interrupt can only come 55 times in a second ... with no additional HW)
.... it is often funny using such a low-level thingies during for exmaple .NET development but helpfull. So I guess the profiler is OK, but during an OS development, the 'tracking' has to be more insightful ...
So:
counters .. ;)
profiler .. hmm
re
Treke |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 67 of 105 | ANN.lu |
Posted by Anonymous on 23-Sep-2003 16:24 GMT | In reply to Comment 13 (Eva): @eva
> ...Absolutely no. Actually Mos is a commercial, mature and usable product ...
?!? |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 68 of 105 | ANN.lu |
Posted by Kolbjørn Barmen on 23-Sep-2003 19:30 GMT | In reply to Comment 64 (Some Farker): > IYO. Every other OS has consistent UI design. The two most successful ones do, as well. Tell me: why *shouldn't* ctrl-v paste in every application? Why *shouldn't* alt activate the pulldown menus?
Because ctrl-v means "the next input shall be escaped" and because "alt" is a qualifier for inserting alternating characters.. hence the name of the key.
Why do people think "the windows way" is "the userfriendly way" when most of it is utter utter nonsense? |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 69 of 105 | ANN.lu |
Posted by hammer on 23-Sep-2003 19:53 GMT | In reply to Comment 38 (Anonymous): Focus on issue at hand NOT on the character. Your statements doesn’t advance your position. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 70 of 105 | ANN.lu |
Posted by Kjetil on 23-Sep-2003 20:42 GMT | In reply to Comment 52 (Anonymous): /* You have heard of a profiler, haven't you? */
Nope, I'm only hobby programmer don't know every thing,
on thing I know getting a profile of system can be good start,
knowing what system calls that are the most frequent is good start,
measuring individual calls and compeering them with other system, can be indication of some thing working or some thing that can be fine tuned more.
(this depends on what the system takes in count, multi user profiles, and fork and clone, treads, memory protection, drivers, etc, so the best OS to compare with should be the most alike, MacOS9 and BeOS should be ideal I think, QNX might be to advanced, and Quark/MorphOS can be also used, how ever you should be testing the different OS's on the same computer, linux can be fine test) |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 71 of 105 | ANN.lu |
Posted by Anonymous on 23-Sep-2003 22:55 GMT | In reply to Comment 70 (Kjetil): >> /* You have heard of a profiler, haven't you? */
> Nope, I'm only hobby programmer don't know every thing
As it happens, it counts the number of entries into functions, so you can forget about the idea to count functions yourself ;-) It also counts the time spent there and outputs that in two ways: the local time (the execution time for the local code in the subroutine) and the total time (adding the execution time of all subroutines that get called there). The profiler prints all that in form of a nice sorted table. Personally, I'm only using it for normal applications. Many components of an OS are not exactly applications but devices, handlers, librariese. No idea if and how those can be profiled, never tried it. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 72 of 105 | ANN.lu |
Posted by Anonymous on 23-Sep-2003 23:31 GMT | In reply to Comment 27 (Fabio Alemagna): >I prefer to judge things by myself, from what I can see, rather than blindly
>believing the parties involved.
It's incredible the arrogance of this guy. I bet you weren't even there, but you prefer to judge things "by yourself" rather than listening to those who are making the OS.
Blindly believing...? You're blind alright.
>I have enough in-dept knowledge to be able to do that.
You have a lot of arrogance, that's for sure. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 73 of 105 | ANN.lu |
Posted by Anonymous on 23-Sep-2003 23:34 GMT | In reply to Comment 41 (Eva): >>> "The first besta of AmigaOs4 will be releasen in 2 months" (Ben Hermann, November 2001)
And if there had not be all the changes that there have been that version of OS4 would have probably been out, but things have taken a different path. So freaking what? |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 74 of 105 | ANN.lu |
Posted by Anonymous on 23-Sep-2003 23:37 GMT | In reply to Comment 53 (Fabio Alemagna): More arrogance from the almighty know-it-all Fabio. Leave it to him to send to hell even one who makes a suggestion on how he should behave (and he really should) |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 75 of 105 | ANN.lu |
Posted by Joe "Floid" Kanowitz on 24-Sep-2003 00:53 GMT | In reply to Comment 58 (Don Cox): Don Cox noted,
> It is a common saying, but IMO consistency is overrated. Programs
> differ so greatly in their purposes that it makes more sense to
> give each the best GUI for the job, especially when they run on
> separate screens.
The examples are, of course, good. But we've got to be careful about generalizations in general. It's not hard to step back and apply some good old fashioned "Zen" to the problem - consistency (and the avoidance of constant wheel-reinvention, and/or retraining for The New Square Wheel) is good; having interfaces optimized for the task or application is equally good.
It's always going to be a back and forth. You can never quite 'solve' the problem, because consistency (use of existing, familiar elements) will always be at odds with doing something new (that the authors of the existing elements didn't think of)... but you can ammeliorate it by recognizing the problem and planning for it. (Contrast, as noted, the ease and 'consistency' of switching between 'inconsistent' full screen programs on the Amiga versus Mac or Windows, which didn't quite plan for it. Turns out 90% of full-screen programs 'don't mind' if you provide a draggable titlebar, and a good number actually still use it to be 'consistent' with their menus.)
At the end of the day, authors will always end up doing things in screwy ways, and we'll always be sitting around arguing over what we will/won't put up with.* If an OS is to have one 'system UI' at all (Workbench/Intuition), its authors are responsible for providing good consistent elements to try to 'talk developers out of' making everything it's own little skinned/special mess, but equally, reducing the 'collateral damage' when someone does want to see if they can do better on their own, rather than pretending it's never an issue. ("Okay, fine, go ahead - just please do it on your own screen... We'll even throw in a consistent titlebar for free unless you *really* can't live with that.")
*Obviously we allow different degrees of inconsistency depending how much we care about the program. People learn [Photoshop, Mozilla, Word, OpenOffice, Lightwave] because they're "big, important apps," - or more accurately, because the value of what they can do seems to outweigh the value of the time wasted learning them. When it's something that seems 'trivial,' like an eRecipeDatabase2000, or one of those transfer utilities for a digital camera or MP3 player that only exists because they were too dumb to make the device fit the USB Mass Storage standard in the first place, you usually don't want to put up with the developer's idea of how much better scrollbars would be if they were green and looked like faucet knobs.
> I don't think it matters at all if some programs are MUI and others
> are Gadtools or Reaction. Nor does it matter if they have different
> colour schemes, or different shaped arrows on the scroll bars, or
> different fonts on screen - any more than it matters if two books on
> a shelf have different typefaces and layouts.
It matters just as far as we think consistency can make things 'easier' or 'more useful.' No further, and no less. ;)
You probably *would* be annoyed if the one author you wanted to read decided it would be better to typeset in Inuit. But you might equally learn to read it if the text was really that valuable to you.
--
Meanwhile, on this MUI vs. Reaction and associated performance business... It's equally "obvious." Two different approaches. Different tradeoffs for each. Personally, I'm finding this Reaction work a little more interesting, because I think we computer nuts in general have completely forgotten how to even *try* to shoot for consistency (and because, as proven, hardware is always going to scale, no matter how slow or 'awful' an idea might be), but both obviously get the job done one way or another, and if one proves to hamper development of 'useful things' (like browser renderers that can take advantage of system font renderers, rather than going, say, the route Links took), then that'll be something to remember for "next time." Same goes if one ends up breaking other apps when a single one wants to introduce a custom whatever.
That's the smell of progress! |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 76 of 105 | ANN.lu |
Posted by hooligan/dcs on 24-Sep-2003 01:39 GMT | Now that things have cooled down and everything that has to be said about the issue is said, how about a short summary what happened. Shall we start from the start, the day when the first report from Italy hit the forums.
1)
We have a report. It clearly states that OS4 on A1 is slow and not ready to be demoed for public. BUT, he didn't know that OS4 was a quick hack put together so it could be shown there. He reported what he saw.
He was correct. What happened here?
He was marked as being a troll, MOS-fanatic, even suggested that he was "no-one" but somebody faking under this name to bring OS4 to bad light.
2)
A new thread, and we get to see videos. Now everybody can see with their own eyes the first report wasn't bs. It was as true as it gets. In this thread the reasons are also explained.
What happened?
Frodon was immediately blamed for running for his own agenda (in this case again, discredit OS4).
Fact is people wanted videos and pictures of OS4 in progress. Thats what they got, this was what was demoed there. It doesn't matter if the videos were shot by Frodon or someone from Hyperion: they would have still showed the exact same thing.
3)
Yet another report. Backing up what was in the first report and what we saw on videos. A very good report with lots of time spent in writing it. At this point everything was already explained and very clear.
What happened?
He was told to be a troll and running also for his own agenda, being a Pegasos user and all.
Fact was, he wasn't even a Pegasos owner.. last time used loong way back.. a year ago.
Everyone can draw their own conclusions inside their own heads. But I tell you this: "blue camp" has always been said to be the worst of all.. this shows the other camp ain't any better.
any comments? |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 77 of 105 | ANN.lu |
Posted by Ben Hermans/Hyperion on 24-Sep-2003 02:38 GMT | In reply to Comment 76 (hooligan/dcs): Any comments?
Small recap:
1. L2 cache not enabled
2. debuglevel erroneously set to 10 instead of 0
3. Smartrefresh not enabled
4. graphics subsystem and graphics driver in 68K code
At least points 1-3 can easily be cured and we're working on 4. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 78 of 105 | ANN.lu |
Posted by Michal Schulz on 24-Sep-2003 04:18 GMT | In reply to Comment 66 (Treke): "(just try to measure an operation in microseconds on a PC, where the timer interrupt can only come 55 times in a second ... with no additional HW)"
Who has told you that? One of PC's timer works with 11931850Hz dividable by value from 65535 to 2. Other (Real Time Clock) generates max. 8192 interrupts pro secod. The APIC timer can generate interrupts with resolution of bus clock (100MHz, 200MHz).
Someone has lied to you with this 55Hz timer ;) |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 79 of 105 | ANN.lu |
Posted by Jon on 24-Sep-2003 04:37 GMT | In reply to Comment 71 (Anonymous): There is at least one open source profiler project for Amiga, AProf. There was also one possible profiler described in the Gamedev.net article series called Enguinity. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 80 of 105 | ANN.lu |
Posted by strobe on 24-Sep-2003 05:02 GMT | In reply to Comment 25 (Hans-Joerg Frieden): I think most people were saying "AOS4 isn't going to run on the A1 any time soon" almost a year ago. You know, when people were buying A1SEs thinking AOS4 was going to be released before Christmas...
I certainly feel I was proven right, and right again. When AOS4 isn't released this year I'll be proven right a third time.
I have nothing against AOS4, but I feel sorry for people who have bought an A1 thinking they would have an Amiga by now. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 81 of 105 | ANN.lu |
Posted by hooligan/dcs on 24-Sep-2003 05:05 GMT | In reply to Comment 77 (Ben Hermans/Hyperion): Yep, these were the reasons I was talking about. I haven't got much understanding for people who were _still_ blaming OS4 was slow after these facts were known (in this case lots of MOS supporters.. so I am not defending the blue brigade either) |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 82 of 105 | ANN.lu |
Posted by Fabio Alemagna on 24-Sep-2003 06:22 GMT | In reply to Comment 72 (Anonymous): > > I prefer to judge things by myself, from what I can see, rather than blindly
> > believing the parties involved.
> It's incredible the arrogance of this guy. I bet you weren't even there, but
> you prefer to judge things "by yourself" rather than listening to those who
> are making the OS.
If I'm able to, I don't see why I shouldn't. Is it a crime to be able to do something?
> Blindly believing...? You're blind alright.
Funny, because The Hyperion guys said I'm right. And it's even more funny, because no even Georg Steger was there, and he hasn't even watched the movies when the made his comment about opaque window moving and, guess what, he was spot on. You know, some people just "can", other - probably like you - can't.
> > I have enough in-dept knowledge to be able to do that.
> You have a lot of arrogance, that's for sure.
If you prefer to equate knowledge to arrogance that's your choice and indeed I'm arrogant, but at least I'm not an anonymous coward which, if you ask me, is 10 times worse. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 83 of 105 | ANN.lu |
Posted by Andrea Maniero on 24-Sep-2003 07:48 GMT | In reply to Comment 77 (Ben Hermans/Hyperion): Small recap:
@Ben Hermans
> 1. L2 cache not enabled
>
> 2. debuglevel erroneously set to 10 instead of 0
>
> 3. Smartrefresh not enabled
>
> 4. graphics subsystem and graphics driver in 68K code
>
> At least points 1-3 can easily be cured and we're working on 4.
Hello! Point 2 was already cured on Sunday, when both me and the author of that other report went to the show. Anyway, I clearly stated that one couldn't expect anything more from the A1 demo: and in fact I believe it was an achievement, and I'm quite happy for having been one of the first people in the world to have the possibility to take a look at it!
*But* what I was really disappointed by was the CSPPC version. Someone said we had too big and unfounded expectations. I don't think so. I distinctly remember that last year you said you wouldn't be present at the Aachen show because you (and in particular the Frieden bros) were too busy with the "last bits" of AOS4. That gave life to a lot of expectations - people on amiga.org were betting for a pre-Christmas release back then. I did believe you, so I was expecting AOS4 to be in a quite different form a year later, on CSPPC. You (and those betatesters that kept saying for the last year how usable the OS was) created a lot of expectations.
Therefore I (and the others who wrote other reports as well) expressed my genuine disappointment and disillusionment, because I feel somewhat cheated. The fact that I am ATM more interested in Peg+MOS (which I didn't hide, because I wanted to be as transparent as I could), doesn't change that.
I really do think you need to improve your straightforwardness (sp?) towards the community.
I honestly didn't want to give your project a cheap shot. I want to point out that I'm sorry for the mess a couple of genuine reports caused. Including the fact that an AOS4 betatester has been kicked out for apparently no reason but trying to defend one member of the community.
Kind regards,
Andrea |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 84 of 105 | ANN.lu |
Posted by Don Cox on 24-Sep-2003 07:52 GMT | In reply to Comment 64 (Some Farker): "Every other OS has consistent UI design. The two most successful ones do, as well. Tell me: why *shouldn't* ctrl-v paste in every application? "
That is not part of the GUI.
"Why *shouldn't* alt activate the pulldown menus?"
Does it? I use a mouse button. What do you mean by "activate" the menus?
"I don't think it matters at all if some programs are MUI and others are Gadtools or Reaction. Nor does it matter if they have different colour schemes, or different shaped arrows on the scroll bars, or different fonts on screen - any more than it matters if two books on a shelf have different typefaces and layouts."
There used to be a UI Hall of Shame on the web; I wish I could find the URL and post it. You just might change your mind if you saw it."
I am talking about having the best possible GUI for each program, not about bad GUIs. A word processor needs a completely different interface from a paint program. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 85 of 105 | ANN.lu |
Posted by Treke on 24-Sep-2003 09:00 GMT | In reply to Comment 78 (Michal Schulz): @Michal Schulz
> One of PC's timer works with 11931850Hz dividable by value from 65535 to 2.
Hmmm, it can be set up using Win32 API ? Managing such a precise timer list will not bring an OS down performance-wise ??
> Someone has lied to you with this 55Hz timer ;)
It seems ;) Thx
re
Treke |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 86 of 105 | ANN.lu |
Posted by Michal Schulz on 24-Sep-2003 09:46 GMT | In reply to Comment 85 (Treke): >Hmmm, it can be set up using Win32 API ?
Don't ask me about Win32 API because I don't know it well enough to judge. :) But yes, precise timer may be used under linux |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 87 of 105 | ANN.lu |
Posted by Kjetil on 24-Sep-2003 10:29 GMT | In reply to Comment 84 (Don Cox): Well I think the “CTRL + V” is part of the GUI, I can’t understand how the “CTRL +V” shortcut know what character it should start copying from and what character it stop copying form, if the gadget provide the information on where the text is located, and what market area of the text to copy?
As for home made GUI vs official GUI, I will not take side on that one, mainly due to this things
.1. It takes longer to design your own gui, some times it better go for the official GUI
.2. Self designed GUI can be better designed for your task
.3. There is lots of crappy self designed GUI’s that should based on official GUI’s as the aliment and scaling is all wrong…
.4. Self designed GUI’s are smaller and can be easier to port.
Clearly there is no black and white on this subject!!! |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 88 of 105 | ANN.lu |
Posted by Bernie Meyer on 24-Sep-2003 11:13 GMT | In reply to Comment 62 (BrianK): While you are of course right that app-specific interfaces increase learning time, I cannot really agree that that is a bad thing....For Grandma, it's a horrific idea, sure. As well as for me where, say, a do-your-income-tax program is concerned. That's because I do that maybe once a year, so I am perpetually at the bottom of the learning curve (just like most Grandmas tend to be, whose computer use is casual at best).However, in a normal day, I spend about 10-12 hours in front of a computer. That happens when you are employed as a software developer. If it takes me 10 hours to learn the interface of a new tool, which will then gives me a timesaving of 5% afterwards, I break even after about a month, and am way ahead after two months.An analogy --- when I build a shelf, I use a jigsaw, a screwdriver and a clawhammer. When a carpenter builds a shelf, he uses professianal rotary saws, a powerdrill and a nailgun. All of those tools are harder for the beginner to master (and, in fact, are physically dangerous) --- but they are the tools of the trade, and not investing the time to learn them would be stupid for somebody who uses that stuff every day.At some point, many (MANY!) years back, I spent a whole weekend adding some functionality for thread-based autokill to my newsreader. Ever since then, my newsreader has done *exactly* what I wanted, and I probably save a couple of minutes each time I read news. So once again, I am actually ahead....Moral of the story --- use the right tool for the job. And what the right tool is depends, among many other things, on how often you have to do the job. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 89 of 105 | ANN.lu |
Posted by Bernie Meyer on 24-Sep-2003 11:17 GMT | In reply to Comment 66 (Treke): > (just try to measure an> operation in microseconds on a PC, where the timer interrupt can only> come 55 times in a second ... with no additional HW)Easy --- you use the RDTSC instruction both when you want to start and stop measuring, and then subtract one from the other. The result is the time that has elapsed in between, with an accuracy of about 10-20 nanoseconds on current hardware. That instruction isn't even privileged, either. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 90 of 105 | ANN.lu |
Posted by dammy on 24-Sep-2003 11:57 GMT | In reply to Comment 74 (Anonymous): > More arrogance from the almighty know-it-all Fabio. Leave it to him to
> send to hell even one who makes a suggestion on how he should behave
> (and he really should)
Oh come on now, if he was really that arrogant son of a bitch, then why isn't he working for Hyperion?
Dammy |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 91 of 105 | ANN.lu |
Posted by Anonymous on 24-Sep-2003 12:13 GMT | In reply to Comment 90 (dammy): There you go,you just said that whoever works for Hyperion is an arrogant son of a bitch. That says a lot about you. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 92 of 105 | ANN.lu |
Posted by Nowee on 24-Sep-2003 12:26 GMT | I think someone that's involved in AROS developpement is fairly well
placed to comment two 'concurrent' Os, that is Mos and Aos. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 93 of 105 | ANN.lu |
Posted by Don Cox on 24-Sep-2003 12:29 GMT | In reply to Comment 87 (Kjetil): "Well I think the “CTRL + V” is part of the GUI, I can’t understand how the “CTRL +V” shortcut know what character it should start copying from and what character it stop copying form, if the gadget provide the information on where the text is located, and what market area of the text to copy?"
Selecting the data is done with the GUI. Dragging it to a new place would be done with the GUI. Marking an insertion point is done with the GUI.
But CTRL-V (or as it should be, Windows-V) is a keyboard operation and can be done without looking at the screen. The GUI equivalent is clicking on a little picture of a roller. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 94 of 105 | ANN.lu |
Posted by Bernie Meyer on 24-Sep-2003 12:34 GMT | In reply to Comment 91 (Anonymous): Actually, he said that whoever is a son of a bitch should be expected to work for Hyperion, not the other way around.Analogy: "If Bernie really was driving a 1975 Beetle, why doesn't he have a garage?". That implies that anyone with a 1975 Beetle is expected to have a garage; It does *not* say that anyone with a garage is expected to drive a 1975 Beetle. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 95 of 105 | ANN.lu |
Posted by MarkTime on 24-Sep-2003 13:12 GMT | In reply to Comment 83 (Andrea Maniero): Andread said:
>>*But* what I was really disappointed by was the CSPPC version
I have a very strong feeling that the CSPPC version will never be a product anyway. If you are hoping fora CSPPC version ,the good news is Hyperion still claims they will finish one.
But this reminds me of the time just before Eyetech cancelled the AmigaONE 1200 project...you could still hope for one, behind the scenes they were not working on it anymore, and publicly they were still claiming it would be released and anyone who said otherwise is an ignorant troll....
but eventually Eyetech realized it wasn't going to happen and cancelled it.
I feel strongly, that while Hyperion hasn't caught up with reality yet, they will someday, and they will cancel this project.
It's too much work, its a dead end, it won't result in profits for the company compared to the work on AmigaONE.
Remember, Hyperion...is a business, and you can tell as they've delayed OS 4 every single time contract work (writing the boot code for AmigaONE ,being one)...has come up. This detour today, isn't the first detour, they have a history of detours. Of course, whenever they have a chance to get some money, they have no choice, being in business for money, but to capitalize on it.
The same is true of the CSPPC project, it doesn't make sense anymore. It will be cancelled, imho. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 96 of 105 | ANN.lu |
Posted by Kjetil on 24-Sep-2003 13:29 GMT | In reply to Comment 95 (MarkTime): I think the CSPPC version have chance, the BPPC version we can not be shore about, they sad it where technically problematic, so get it to work so we can only, wonder if the work load equivalent to the money input or not,
I think the CSPPC version will be cut down version of AmigaOS4.0, I think you get what runs on both AmigaOne and Amiga4000 and every thing else is drooped if it requires any effort to adapt, basically AmigaOS4.0 with out the snap drivers.
I think they will analyze the publics interest after the Amiga4000 version and AmigaOne XE,SE versions are on the market, I might even be that AmigaOne Lite will on heir priority list then the Blizzard version.
There are more Amiga1200 out there the Amiga4000 how many users will be interested/using in the Amiga1200 after the AmigaOne's are on the market? |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 97 of 105 | ANN.lu |
Posted by Kjetil on 24-Sep-2003 13:39 GMT | In reply to Comment 93 (Don Cox): “CTRL +V” shortcut need to know what gadget that is active to copy the market text from it? or are suggest it should analyze the desktop image for characters and decode the characters with a black areas from gfx to text? sound like CPU heavy task if you ask me. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 98 of 105 | ANN.lu |
Posted by BrianK on 24-Sep-2003 13:57 GMT | In reply to Comment 88 (Bernie Meyer): The right tool for the right job is typically preferred. So, if I learn File --- Save that's the 'tool' for saving within an application. If I learn to cut with ctrl-v that's the 'tool' for cutting.
The problem is with the consistency problem. If I want to cut and paste in one application why should I need to learn a different 'tool' (perhaps Amiga-Z) to cut? If I want to save why can I not use my file-save 'tool'? There are common pieces within applications that certain learned shortcuts and methods should be kept consistent.
Often various people have their own idea of what a great interface would be. Instead to the average joe PC using public they aren't going to deal with the hassle to learn a new interface.
Consistency of an interface helps to reduce learning but yes it could reduce usefulness overall compared to a different method. There's always a 'better' way to do something and your better way could be less useful for me. Imagine the difficulty in creating a system where the end user creates the interface for each program to suit their liking? |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 99 of 105 | ANN.lu |
Posted by Jorge W on 24-Sep-2003 23:52 GMT | I only can say this:
for (int i=0;i=1000000000;++i)
{
}
is faster than:
for (int i=0;i=1000000000;++)
{
cout << "error on" << printDebug(debugCode);
if (functionBlah)
{
_functionBloh();
}
if (functionBlih)
{
_functionBluh();
}
// lotsa more if...thens follow
...
}
Notice: Regarding the performance on Saturday, please note that the AmigaOne had the debuglevel set to 10 and also that the graphics system still ran interpreted. |
|
AOS4 on AOne slow? Perhaps, perhaps not. : Comment 100 of 105 | ANN.lu |
Posted by Joe "Floid" Kanowitz on 25-Sep-2003 03:22 GMT | In reply to Comment 84 (Don Cox): [Sat on this one for a day or two, decided to post it anyway. Hope I didn't leave any sentences hanging.]
Don Cox said,
> > [Someone else's ^V = Paste argument.]
> That is not part of the GUI.
It's part of the 'interface,' or as some people like to put it, 'interaction' as a whole. Though right now we're arguing GUIs mostly as relate to speed of rendering/flexibility of updates around other events (window drags, etc), so I'm not sure how this came up.
> > "Why *shouldn't* alt activate the pulldown menus?"
> Does it? I use a mouse button. What do you mean by "activate" the menus?
I think he's talking about what I consider the "TurboVision" approach to mouse/keyboard equality that a lot of companies hit on with MS-DOS (where mice were rare), which is now part of Windows and practically every other WIMPstrosity except Mac. It's nice to get at pull-downs from the keyboard, and *if the hotkeys are organized logically,* it can be a self-documenting way to gain keyboard proficiency. I can never remember if the Amiga had/has any way to expose or select the menus from the keyboard (beyond using the mouse emulation), as others whined to me when I brought up the concept, it'd be a bit fiddly to implement in the classic UI because they're hidden all the time.
But knowing that the RMB will work the proverbial 96% of the time is, again, consistency.
> I am talking about having the best possible GUI for each program, not
> about bad GUIs. A word processor needs a completely different interface
> from a paint program.
Unless you're on a Newton, in which case they probably both look like blank pages. (On an unrelated note, this is actually where we get totally screwed; say you want DTP, do you use separate word processing and paint apps? Do you use an integrated tool, or a suite? If it's pay software, what happens when good monolithic apps are cost-prohibitive versus one app that does everything badly and reinvents a truck fleet of wheels? Do you use something like OpenDoc, where everything's a mess of objects and you can never be sure quite how it'll all interact? Okay, you've found something that works for you, now what do you tell newbies to do? ... and so on.)
---
For everyone else: This thread didn't suck while it was still technical. In open-sourceland, where people have fewer monetary interests, it's called "discussion" and it happens all the time. Obviously backwards Commies, those guys use it *mainly* to clarify understandings and develop insight, rather than our scene's enlightened pheremonal marking.
Props to both Fabio and Hans-Joerg, at least before anyone gave in to the baiting. |
|
Anonymous, there are 105 items in your selection (but only 55 shown due to limitation) [1 - 50] [51 - 100] [101 - 105] |
|
- User Menu
-
- About ANN archives
- The ANN archives is powered by #AmigaZeux. It was updated daily (news last: 22-Oct-2004; comments last: 18-May-2005).
ANN.lu was created, previously owned and maintained by Christian Kemp, www.ckemp.com.
- Contribute
- Not possible at this time!
- Search ANN archives
- Advanced search
- Hosting
- ANN.lu was hosted by Dreamhost. Sign up through this link, mention "ckemp" as referrer and he will get a 10% commission on any account you purchase.
Please show your appreciation for any past, present and future work on ANN.lu by making a contribution via PayPal.
|