| [Forum] Try only to realise the truth | ANN.lu |
|
| Try only to realise the truth : Comment 101 of 335 | ANN.lu |
| Posted by Anonymous on 05-Jul-2004 04:14 GMT | In reply to Comment 37 (reflect):
Egomaniacs like to believe their own truths, Alemagna is one of them. |
|
| Try only to realise the truth : Comment 102 of 335 | ANN.lu |
| Posted by priest on 05-Jul-2004 04:53 GMT | In reply to Comment 35 (Fabio Alemagna):
"But didn't Hyperion already say very long ago, way before the prerelease was sent to customers, that DMA was working?"
Right. IDE DMA test application was working 100%. (HW banging, no drivers)
IIRC:
IDE DMA driver reached beta state after prerelease CD already "went gold". It took 2-3 months to deliver the CD to customers. During that time IDE DMA driver was polished/bug fixed and went into beta testing. Betatesters have reported "no DMA errors". |
|
| Try only to realise the truth : Comment 103 of 335 | ANN.lu |
| Posted by priest on 05-Jul-2004 05:06 GMT | In reply to Comment 67 (Fabio Alemagna):
"Articia is buggy, in that it doesn't fulfill its specifications"
It depends. I think Articia pretty much fullfills its specifications.
But those specification might not comply 100% with PCI standard. etc. ;-) |
|
| Try only to realise the truth : Comment 104 of 335 | ANN.lu |
| Posted by Anonymous on 05-Jul-2004 06:02 GMT | In reply to Comment 58 (Thomas Frieden):
>12 seconds. Really laughable.
sure that's not bad, but compared to MorphOS which is about 7-8 s on G3 with full USB support... |
|
| Try only to realise the truth : Comment 105 of 335 | ANN.lu |
| Posted by For Thomas Friend, Hyperion, Eytech and co on 05-Jul-2004 06:08 GMT | In reply to Comment 62 (Thomas Frieden):
"The answer is quite simple. The DMA driver was very new at the time of the prerelease. Naturally, new software is risky to release. Since we know the DMA issue is kind of sensitive, we wanted to be absolutely sure that there is no bug in a1ide.device. Otherwise, all those nice people like you would jump to conclusion and shout "The Articia S is buggy" all over the place, regardless of whether it's a software error or not. People like you would not care, nor ask.
Therefore, we decided not to release the DMA driver. It's as simple as that, and I think it's a sensible approach.
Happy flaming."
You misunderstand.
The point is here is not to discuss about the "possibility" to make a DMA drvier without cache coherency.
The point is to show that the ArticiaS is crap unless that you, Hyperion and Eyetech trying to say.
Will you argue that no cache coherency is a feature ? Because you have to flush everything by yourself ?
And what about the SE one, this one will never work. Even if you do all that is needed to don't trust Cache coherency.
I'm really curious to have you anwser. |
|
| Try only to realise the truth : Comment 106 of 335 | ANN.lu |
|
| Try only to realise the truth : Comment 107 of 335 | ANN.lu |
| Posted by Sammy Nordström on 05-Jul-2004 06:24 GMT | In reply to Comment 50 (Emeric SH):
>"Fact: 3rd party beta-testers confirm DMA is working properly with AmigaOS4."
>
>There are hundreds of AmigaONE users out there using their AmigaONE without
>problems.
Surely you must see the difference between "using" and "beta testing"? |
|
| Try only to realise the truth : Comment 108 of 335 | ANN.lu |
| Posted by cheesegrate on 05-Jul-2004 07:01 GMT | In reply to Comment 99 (James Carroll):
hey carrol, so i guess u agree with nerp that ann sucks..
so what are u doing here, shouldn't u be masturbating to fleecy's picture? oh wait... u are. |
|
| Try only to realise the truth : Comment 109 of 335 | ANN.lu |
| Posted by NeRP on 05-Jul-2004 07:15 GMT | In reply to Comment 108 (cheesegrate):
Sorry, but what does fleecy have to do with anything, you (not "u") useless ass? |
|
| Try only to realise the truth : Comment 110 of 335 | ANN.lu |
|
| Try only to realise the truth : Comment 111 of 335 | ANN.lu |
| Posted by uhm on 05-Jul-2004 07:37 GMT | In reply to Comment 110 (Kolbjørn Barmen):
Care to elaborate? |
|
| Try only to realise the truth : Comment 112 of 335 | ANN.lu |
| Posted by James Carroll on 05-Jul-2004 07:45 GMT | In reply to Comment 108 (cheesegrate):
>hey carrol, so i guess u agree with nerp that ann sucks..
Never said that, but thanks for putting words in my mouth.
>so what are u doing here, shouldn't u be masturbating to fleecy's picture? oh >wait... u are.
Your insults need work. |
|
| Try only to realise the truth : Comment 113 of 335 | ANN.lu |
| Posted by Metalheart on 05-Jul-2004 08:01 GMT | In reply to Comment 20 (Darth_X):
do you ? |
|
| Try only to realise the truth : Comment 114 of 335 | ANN.lu |
| Posted by Richi on 05-Jul-2004 08:03 GMT | In reply to Comment 108 (cheesegrate):
Where is the moderator?! |
|
| Try only to realise the truth : Comment 115 of 335 | ANN.lu |
| Posted by Emeric SH on 05-Jul-2004 08:20 GMT | In reply to Comment 107 (Sammy Nordström):
"Surely you must see the difference between "using" and "beta testing"?"
Interestingly, "ordinary" users and not betatesters found and proved the DMA bug existance. The same happened to the bug in IBrowse demo on OS4...
It may be a FACT if you like that OS4 betatesters didn't encountered problems, but that hardly proves anything. |
|
| Try only to realise the truth : Comment 116 of 335 | ANN.lu |
| Posted by Sammy Nordström on 05-Jul-2004 08:34 GMT | In reply to Comment 115 (Emeric SH):
>Interestingly, "ordinary" users and not betatesters found and proved the DMA
>bug existance.
That's funny, I thought it has yet to be proved that it is a bug rather than a hardware compatibility issue. Furthermore, I thought AmigaOS4 hasn't been available to "ordinary users" until the recent release of the AmigaOS4 developer pre-release and that this issue has been known and discussed openly for quite some time prior to the AmigaOS4 developer pre-release. But then, by all means, show me the "proof" those "ordinary users" came up with. |
|
| Try only to realise the truth : Comment 117 of 335 | ANN.lu |
| Posted by Emeric SH on 05-Jul-2004 08:40 GMT | >>Interestingly, "ordinary" users and not betatesters found and proved the DMA
>>bug existance.
>That's funny, I thought it has yet to be proved that it is a bug rather than a >hardware compatibility issue.
Read my words more carefully then :) Whether it is a "hardware compatibility issue", a "bug in ArticiaS", a "feature of ArticiaS", "a Linux driver bug", it's a bug in DMA, till it corrupts the transfer, regardless of it's origin.
Proof to it's existance were simple screenshots showing corrupt MD5 sums after transferring the same data over and over with DMA enabled, just dig deep enough in ANN archives. |
|
| Try only to realise the truth : Comment 118 of 335 | ANN.lu |
| Posted by James Carroll on 05-Jul-2004 08:44 GMT | In reply to Comment 117 (Emeric SH):
davep showed a screenshot of his a1 running a dma enabled driver, complete with checksums etc awhile ago.. theres definately a problem with dma, but not on os4 it seems. just linux. (not trying to flame) |
|
| Try only to realise the truth : Comment 119 of 335 | ANN.lu |
| Posted by Emeric SH on 05-Jul-2004 08:50 GMT | In reply to Comment 118 (James Carroll):
We've heard from DaveP that there are no DMA problems in Linux as well. |
|
| Try only to realise the truth : Comment 120 of 335 | ANN.lu |
| Posted by Sammy Nordström on 05-Jul-2004 08:57 GMT | In reply to Comment 117 (Emeric SH):
The so-called proof you are refering to was from a LinuxPPC user and it has been known for quite some time now that DMA on the AmigaOne doesn't work in LinuxPPC. Wether it is a bug or not is determined by each individuals subjective point of view on how you define a bug but also what definition that is most suitable for your own personal views.
Furthermore, even if it is a "bug", we have yet to determine wether it's in the hardware or the software.
In other words; the statement "there is a bug in the DMA" is based on a multitude of assumptions that has yet to be established as an objective reality. |
|
| Try only to realise the truth : Comment 121 of 335 | ANN.lu |
| Posted by James Carroll on 05-Jul-2004 08:59 GMT | In reply to Comment 119 (Emeric SH):
Well he was wrong. He's only human. But he conducted an extensive soak test on OS4 which showed no problems. |
|
| Try only to realise the truth : Comment 122 of 335 | ANN.lu |
| Posted by Emeric SH on 05-Jul-2004 09:01 GMT | In reply to Comment 121 (James Carroll):
"Well he was wrong. He's only human. But he conducted an extensive soak test on OS4 which showed no problems."
Like his Linux stress test I presume. |
|
| Try only to realise the truth : Comment 123 of 335 | ANN.lu |
| Posted by Fabio Alemagna on 05-Jul-2004 09:03 GMT | In reply to Comment 96 (NeRP):
> I guess all A3000 owners were also "fooled" when they found out that their
> zorro3 slots didn't work. Then they had to pay $100US or more for a new revision
> of the buster chip that still didn't work. Then they had to buy Rev11 of the
> buster chip to finally get working zorro3.
Yup, you guessed right, they were indeed fooled. Not only that, they were fool enough to pay more momeny to get the thing repaired, when the guarantee should have covered it all. |
|
| Try only to realise the truth : Comment 124 of 335 | ANN.lu |
| Posted by Emeric SH on 05-Jul-2004 09:06 GMT | "In other words; the statement "there is a bug in the DMA" is based on a multitude of assumptions that has yet to be established as an objective reality."
Sorry. If I transfer a file, and the result is corrupted I usually call it a bug. Whether it is in Articia, driver, whatever,
Have you heard "Keine bug, feature?" |
|
| Try only to realise the truth : Comment 125 of 335 | ANN.lu |
| Posted by Fabio Alemagna on 05-Jul-2004 09:07 GMT | In reply to Comment 102 (priest):
> Right. IDE DMA test application was working 100%. (HW banging, no drivers)
Oh, I see. In other words, they couldn't even be certain it would work when the driver was made and all pieces had to interact with each other. |
|
| Try only to realise the truth : Comment 126 of 335 | ANN.lu |
|
| Try only to realise the truth : Comment 127 of 335 | ANN.lu |
| Posted by Fabio Alemagna on 05-Jul-2004 09:10 GMT | In reply to Comment 107 (Sammy Nordström):
> Surely you must see the difference between "using" and "beta testing"?
One'd expect that betatesting is a way of testing something that can highlight bugs better than normal usage... |
|
| Try only to realise the truth : Comment 128 of 335 | ANN.lu |
| Posted by Bernie Meyer on 05-Jul-2004 09:12 GMT | In reply to Comment 115 (Emeric SH):
Interesting anecdote --- how long do you think it took, *after* the release of Amithlon, for some "ordinary user" to notice that hitting the right (as opposed to left) control key on their PC keyboard would result in all sorts of weird behaviour?How long, on the other hand, did H&P run a "beta test" on the stuff *before* shipping it?Say what you want, but there is no substitute to having real people beat up "your baby" if you want to know whether there are problems somewhere :) |
|
| Try only to realise the truth : Comment 129 of 335 | ANN.lu |
| Posted by Anonymous on 05-Jul-2004 09:12 GMT | In reply to Comment 120 (Sammy Nordström):
I see your point.
Non reliable cache coherency support is a feature. humm ... ! |
|
| Try only to realise the truth : Comment 130 of 335 | ANN.lu |
| Posted by Anonymous on 05-Jul-2004 09:18 GMT | In reply to Comment 120 (Sammy Nordström):
And how do you explainthat the very same driver works well on Pegasos I with April ?? |
|
| Try only to realise the truth : Comment 131 of 335 | ANN.lu |
| Posted by Don Cox on 05-Jul-2004 09:24 GMT | In reply to Comment 127 (Fabio Alemagna):
"One'd expect that betatesting is a way of testing something that can highlight bugs better than normal usage..."
Not so. Beta testing is done by a small (or very small) number of users with a limited range of hardware. All kinds of problems emerge when hundreds or thousands of users run the software on all kinds of hardware. No two computers are identical.
Beta testing picks up only the obvious bugs. Also, beta testers usually understand the program well, so they do not find faults in the documentation. |
|
| Try only to realise the truth : Comment 132 of 335 | ANN.lu |
| Posted by Don Cox on 05-Jul-2004 09:25 GMT | In reply to Comment 130 (Anonymous):
"And how do you explainthat the very same driver works well on Pegasos I with April ??"
Nobody can explain that except the person who designed April 2, as what it does is a secret. |
|
| Try only to realise the truth : Comment 133 of 335 | ANN.lu |
| Posted by Sammy Nordström on 05-Jul-2004 09:27 GMT | In reply to Comment 124 (Emeric SH):
>If I transfer a file, and the result is corrupted I usually call it a bug.
If you call it a bug, you are saying that it is:
a. A defect or difficulty, as in a system or design.
b. Computer Science. A defect in the code or routine of a program.
If that's not what you intend to say, you are obviously not using the term correctly. |
|
| Try only to realise the truth : Comment 134 of 335 | ANN.lu |
| Posted by Fabio Alemagna on 05-Jul-2004 09:36 GMT | In reply to Comment 131 (Don Cox):
> Not so. Beta testing is done by a small (or very small) number of users with a
> limited range of hardware. All kinds of problems emerge when hundreds or
> thousands of users run the software on all kinds of hardware. No two computers
> are identical.
Don, thanks for having explained how not to do betatesting :-)
Seriously, if you rely on normal usage to do discover all kind of bugs, then you're basically making a public betatesting session.
Of course the more people the more bugs get discovered in the same amount of time, however it's also true that the more time you betatest, the more bugs get discovered with the same amount of people, simply because what matters is the total usage time.
We can thus say that AOS4 is in an early betatesting phase, still. Public betatesting, even. |
|
| Try only to realise the truth : Comment 135 of 335 | ANN.lu |
| Posted by Anonymous on 05-Jul-2004 10:01 GMT | Mr. Wøntermiyer will publish information on how to reproduce the bug on OS4
after OS4 final has been shipped. |
|
| Try only to realise the truth : Comment 136 of 335 | ANN.lu |
| Posted by Emeric SH on 05-Jul-2004 10:12 GMT | In reply to Comment 128 (Bernie Meyer):
"Say what you want, but there is no substitute to having real people beat up "your baby" if you want to know whether there are problems somewhere :)"
Absolutely agree :) |
|
| Try only to realise the truth : Comment 137 of 335 | ANN.lu |
| Posted by Emeric SH on 05-Jul-2004 10:24 GMT | In reply to Comment 133 (Sammy Nordström):
"If that's not what you intend to say, you are obviously not using the term correctly."
bug - An unwanted and unintended property of a program or piece of hardware, especially one that causes it to malfunction. Antonym of feature.
Now, trashed data after DMA transfer mets "malfunction" in my eyes.
But for your pleasure, a snippet from http://computing-dictionary.thefreedictionary.com/feature - what is a feature?:
6. A bug that has been documented. To call something a feature sometimes means the author of the program did not consider the particular case, and that the program responded in a way that was unexpected but not strictly incorrect. A standard joke is that a bug can be turned into a feature simply by documenting it (then theoretically no one can complain about it because it's in the manual), or even by simply declaring it to be good. "That's not a bug, that's a feature!" is a common catch-phrase.
Kind of familiar, isn't it? :) |
|
| Try only to realise the truth : Comment 138 of 335 | ANN.lu |
| Posted by priest on 05-Jul-2004 10:27 GMT | In reply to Comment 130 (Anonymous):
ArtisiaS DMA reguires modifications to upper layer of Linux DMA handling (unless April is used). |
|
| Try only to realise the truth : Comment 139 of 335 | ANN.lu |
| Posted by priest on 05-Jul-2004 10:35 GMT | In reply to Comment 137 (Emeric SH):
>"bug - An unwanted and unintended property of a program or piece of hardware, especially one that causes it to malfunction. Antonym of feature."
>
>Now, trashed data after DMA transfer mets "malfunction" in my eyes.
Things get complicated if the HW works with one SW but not with the other.
If AOS4 DMA drivers do not cause trashed data, but linux DMA drivers do. |
|
| Try only to realise the truth : Comment 140 of 335 | ANN.lu |
| Posted by Anonymous on 05-Jul-2004 10:37 GMT | In reply to Comment 138 (priest):
>ArtisiaS DMA reguires modifications to upper layer of Linux DMA handling (unless >April is used).
Do you mean it requiere manual cache flushing/invalidating because the cache coherency is not safe ?
In this case it's already supported by Linux, hack the header to remove CACHE_COHERENCY for PowerPC 60x family (G3/G4 ...). And pray for the driver you are using support it.
Bye |
|
| Try only to realise the truth : Comment 141 of 335 | ANN.lu |
| Posted by Sammy Nordström on 05-Jul-2004 10:55 GMT | In reply to Comment 137 (Emeric SH):
Your own choice of defintition for the word "bug" actually reinforces my argument that the cause for the problem is what defines wether it is a bug or not. You see, both computers and computer programs can malfunction without bugs and wether the DMA related problems on the AmigaOne is caused by a bug or not remains to be proven. |
|
| Try only to realise the truth : Comment 142 of 335 | ANN.lu |
| Posted by priest on 05-Jul-2004 11:09 GMT | In reply to Comment 140 (Anonymous):
Surely some people have tried that already. Not??? |
|
| Try only to realise the truth : Comment 143 of 335 | ANN.lu |
| Posted by priest on 05-Jul-2004 11:10 GMT | In reply to Comment 140 (Anonymous):
btw. what's the performance penalty on x86 systems when using that SW cache coherency? |
|
| Try only to realise the truth : Comment 144 of 335 | ANN.lu |
| Posted by Anonymous on 05-Jul-2004 11:46 GMT | In reply to Comment 143 (priest):
Don't change the subjet like that :-P
Is an unreliable cache coherency support a feature ?
Bye |
|
| Try only to realise the truth : Comment 145 of 335 | ANN.lu |
| Posted by Amon_Re on 05-Jul-2004 12:06 GMT | Did i just step a year back in time? I tought the mindless flaming & fuming at the mouth was in our passed *sigh*
How many times will we have the same stupid flamewars?
When will you lot stop foaming at the mouth? When AOS4 is released with DMA drivers? When Linux finally properly support the A1? When the sky falls down?
What does it take to shut all of you imbecils up?!
Cheers |
|
| Try only to realise the truth : Comment 146 of 335 | ANN.lu |
| Posted by Don Cox on 05-Jul-2004 12:13 GMT | In reply to Comment 134 (Fabio Alemagna):
"Of course the more people the more bugs get discovered in the same amount of time, however it's also true that the more time you betatest, the more bugs get discovered with the same amount of people, simply because what matters is the total usage time."
The range of hardware used is equally important, unless the code is some very simple command-line tool.
The advantage of experienced beta testers over average users is that they are better at describing exactly the circumstances in which a bug appears. |
|
| Try only to realise the truth : Comment 147 of 335 | ANN.lu |
| Posted by Don Cox on 05-Jul-2004 12:16 GMT | In reply to Comment 145 (Amon_Re):
"What does it take to shut all of you imbeciles up?!"
Presumably, independent tests showing DMA working perfectly on both Linux and AOS4 on an AmigaOne. |
|
| Try only to realise the truth : Comment 148 of 335 | ANN.lu |
| Posted by Amon_Re on 05-Jul-2004 12:18 GMT | In reply to Comment 147 (Don Cox):
I'm willing to bet some people would still not be satisfied then |
|
| Try only to realise the truth : Comment 149 of 335 | ANN.lu |
| Posted by priest on 05-Jul-2004 12:22 GMT | In reply to Comment 144 (Anonymous):
" Is an unreliable cache coherency support a feature ? "
I do not know what it is untill the OS is updated/written to TRY to make it reliable, it is being done for AOS4. (apparently not for Linux and not for Linux/OpenBSD/MOS + peg1 without April)
Then it perhaps becomes a "SW cache coherency" -feature on AOS4 but a unsupported cache coherency feature (or "unreliable cache coherency" feature) on Linux (and on peg1 without April). ;-)
BUT, if it still fails, it might morph to a "SW bug to be fixed", etc etc... untill SW developers give up (-> it's a HW bug) or untill no one any more finds glitches in data transfer ("SW cache coherency" -feature).
(if the functionality is result of MAI's attempt to gain some new advantage over the competition, it would have been wise to make it optional so that traditional HW cache coherency would work... but nowdays it seems that MAI was just not too smart in general...) |
|
| Try only to realise the truth : Comment 150 of 335 | ANN.lu |
| Posted by priest on 05-Jul-2004 12:24 GMT | In reply to Comment 147 (Don Cox):
I do not think so. People will then start to nag about the A1 DMA issues of NetBSD, OpenBSD, PPC AROS, etc... |
|
|
- 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.
|
