“I read with some interest about the GNOME startup bounty. As Stephen O’Grady pointed out, this problem is indeed perfect for DTrace. To get a feel for the problem, I wrote a very simple D script.” Read more…
“I read with some interest about the GNOME startup bounty. As Stephen O’Grady pointed out, this problem is indeed perfect for DTrace. To get a feel for the problem, I wrote a very simple D script.” Read more…
I know many have voiced concerns that OSNews doesn’t indicate original content from submitted content and while I’ve never really agreed with those who made that claim, this posting really does prove their point.
As I read the news summary, I thought it was Thom Holwerda’s own words that was being posted. Instead, it was the first paragraph of Bryan Cantrill’s blog. This should have been made clear in the original news article. It is rather unprofessional to not do so.
Forgot quote marks, sorry.
Fixed.
PS: We do. We have a policy on OSNews that states that we only use quote marks when directly quoting a person. For normal articles, it is allowed by copyright law to use a small paragraph. Note that it says: “Linked by…” not “Written by…” or “Posted by…”.
why the gnome developers don’t focus on real features needed by the users ?
peoples don’t cares of DTrace
I don’t want to start a flamewar but I have noticed that the startup time of Gnome is very acceptable compared to KDE which takes an aeon to load.
Could you back up what you say with values ? Time, numbers, the amount of processes started, what daemons, applets, get loaded and so on ?
Are these trolls trained or something, they inmediatly start asking proves, simple, start GNOME then start KDE and you would see the difference, no rockets science there. And no, don’t spect the user to know how to edit config files to remove daemonds to make KDE faster, its all in the defaults and that’s the impression it gives.
Are these trolls trained or something, they inmediatly start asking proves, simple, start GNOME then start KDE and you would see the difference, no rockets science there.
Yer, and KDE still starts as fast, if not faster than Gnome. It should be faster though, so it’s a problem both desktops have.
And no, don’t spect the user to know how to edit config files to remove daemonds to make KDE faster, its all in the defaults and that’s the impression it gives.
You can’t even do that with Gnome. It’s all inside.
Yer, I sometimes wonder where these trolls come from and what their point actually is ;-).
Twm is even faster, it loads almost instantly. So, do you plan to use that? More features (and services) means longer startup times.
They’re both still pitifully slow. Blah blah Windows yeah yeah — who cares? Why can’t we just have a fast, clean desktop OS?
Still, glad to see some people ARE paying attention now. That recent discovery of a 300kbyte XML file being loaded and parsed JUST TO SET THE VOLUME showed how abysmal it had all gotten, and how open source doesn’t necessarily mean elegant code and design.
open source doesn’t necessarily mean elegant code and design.
I think the problem is the opposite actually. The problem with open source in many cases is that the developers WANT it to be elegant because you don’t want the world seeing your ugly hackish code. In most cases I’ve seen elgance is a trade off you sometimes make for speed. In the case of Windows I wouldn’t call the registry elegant by any means but it’s sure as hell fast because it’s in a binary format as opposed to flat xml files. Xml is much more difficult to corrupt, human readable, and generally easier to work with. Unfortunately it’s going to take much longer to parse xml than it is to do a lookup in a binary structure of some sort.
I personally think that a lot effort can be put into speeding up the boot process before we even get to X. Loading every single program serially is about as wastefull as you can get.
That recent discovery of a 300kbyte XML file being loaded and parsed JUST TO SET THE VOLUME showed how abysmal it had all gotten, and how open source doesn’t necessarily mean elegant code and design.
You are only looking at the mid-part of the process. Now that the problem has been found, a fix will be made and now you’ll have elegent code and design. You’re looking at a rough draft and condemning it for not being elegent and well organized.
I wish they would have a bounty for removing the sometimes very long logout time. It has happened on about all recent distributions.
Now I have Ubuntu 5.04 installed and yesterday I try to shut down the computer from the gnome logout dialog. I select Shut Down, the dialog disappears and then the system does nothing for about a minute, then X closes and the system begins to shut down.
Far more annoying than the startup time.
PC’s are designed to stay on indefinately, last time I shutdown was when I inserted a new gfx card 1 year ago..
so I suppose it depends.. I would suggest you leave your pc on (unless you using a notebook) It’s really just the monitor that gobbles power..
No they’re not, most pcs are used then turned off, at least for the average user.
I don’t want my pc to be server-like and stay always on.
Why should the user change its habits ? Linux on the desktop should adapt or die.
PC’s are designed to stay on indefinately, last time I shutdown was when I inserted a new gfx card 1 year ago..
so I suppose it depends.. I would suggest you leave your pc on (unless you using a notebook) It’s really just the monitor that gobbles power..
Who said PC’s are designed to stay on indefinitely? You’re making that up. It’s stupid to leave your computer on all the time and doing so wastes electricity and money.
i leave mine on… why shut it down… i have the montior kick off after 15min of usage and usually flick it off at night but thats it…
any ideas which one wears out a electrical device faster – leaving it on or change of states (off/on)? give you one guess at it….
i would say a computer left on would cost you about a few dollars a month…. give or take…
heck i use mine every few hours so why would i shut it down? if you arent going ot use it for a day or so then maybe but otherwise on is better….
oh and saying something is stupid is just stupid!
not realy.
the most damaging thing you can do to the electronics is to repeatidly turn them off and on like shutting down when you are done.
it is called sleep mode: your computer will use very little power so no resources are really wasted.
when does a lighbulb blow? most likely to happen when you switch it on. Why? electricity initialy SURGES into it.
Funny enough leaving a flourescent light on continuously actually is cheaper than switching it on every morning.
Are you using pppoe by chance? When I’m at home and using pppoe, ubuntu takes a rather long time to logout/shutdown (in the tens of seconds), while when I’m connected to a run of the mill dhcp router, there is no such lag. Just a thought.
-Mark Brophy
Who gives a damn how long it takes to load. Shorter is better, sure, but I’ll take a longer load time in exchange for features that mean something any day.
That’s why I don’t mind using GNOME at all.
Does anyone know if specific patches are applied to the login sequence in Ubuntu? I have Hoary running on a P4 and it appears to only take a few seconds from GDM to desktop.
I suppose I do have the luxury of a bucket load of RAM though…
I think loading things in parallel has to be the way to go, as well as mmap caches for the regularly used items – it seems crazy the “x” of close is loaded every time it is used!
It’s great to see SUN cares about GNOME, hopefully SUN and the OpenSolaris community can do for GNOME performance wise what Ubuntu has done usability wise and more! They have the tools that no one else does to do this analysis (just as any Solaris user now does). It’s wonderful…
Yes, more features inevitably means more delays in starting. Nobody’s arguing against that. BUT, when you have a 300 kbyte file being parsed JUST to set the darn volume, something is terribly wrong. And that’s just ONE tiny example — this sort of thing is increasingly common on the Linux desktop.
GNOME, KDE, OpenOffice.org, Mozilla etc. could all be MUCH faster if real effort and care went into clean, trim design and making good use of resources. Sure, GNOME is always going to be slower than TWM, but that’s no excuse for its current bloatness.
I’ve seen many newcomers to Linux turned off by the slowness and weight. We need MANY REAL incentives for people to switch!
“You’re looking at a rough draft”
Eh? How is a supposedly production-ready release, at TWO-POINT-TEN, a “rough draft”? How long has that horrible example of wastage been in there? How much else is there?
This is sloppiness, and not just in 0.1.
Eh? How is a supposedly production-ready release, at TWO-POINT-TEN, a “rough draft”? How long has that horrible example of wastage been in there? How much else is there?
This is sloppiness, and not just in 0.1.
As long as humans are writing software it will never be perfect. Windows, BeOS, Linux, Solaris, etc. *all* have their problems. Stop harping whining and complaining about things. The point is something is being done to analyze the problem so that a fix can be made. Move on, and stop wasting our time with pointless rants.
“As long as humans are writing software it will never be perfect. Windows, BeOS, Linux, Solaris, etc. *all* have their problems. Stop harping whining and complaining about things. The point is something is being done to analyze the problem so that a fix can be made. Move on, and stop wasting our time with pointless rants.”
Perfection isn’t what’s being looked at. Common sense tells you that doing a lot (loading a big file) to do a little (change volume) isn’t a good move. The fact that this has persisted across several versions just reenforces the point. The fact that something is presently being done about the issue is good, but it misses the point. If whomever wrote the software is that sloppy over something that simple. That calls into question other more complicated things. The fact that even other OS’s have issues likewise misses the point. We’re talking about Linux, and it’s issues. It’s Linux that people want to get better. It’s not going to get better if every issue brought up is viewed as “harping whining and complaining”. Is that what you want?
We’re talking about Linux, and it’s issues. It’s Linux that people want to get better. It’s not going to get better if every issue brought up is viewed as “harping whining and complaining”. Is that what you want?
Well said.
Perfection isn’t what’s being looked at. Common sense tells you that doing a lot (loading a big file) to do a little (change volume) isn’t a good move. The fact that this has persisted across several versions just reenforces the point.
*EVERY* OS has the same issues, every OS has at one time or another had something persist across several versions. Get off your high horse and realise that people aren’t perfect and make mistakes. Did you ever stop to think that at one time that file *wasn’t* 300kb? Maybe at one time it was 3kb…
You go on about common sense and then you turn around and show that you certainly don’t have any development common sense. People are human, they make mistakes, what may seem obvious to you is not obvious to everyone. Many people feel the same way about Microsoft’s security track record, or some “obvious” deficiencies in other products. Don’t *assume*.
We’re talking about Linux, and it’s issues. It’s Linux that people want to get better. It’s not going to get better if every issue brought up is viewed as “harping whining and complaining”. Is that what you want?
No, I want you to stop needlessly ranting. Someone look at a problem, finds it to get a better analysis, and goes to fix it, and all you can do is view it as a failure instead of an opportunity for success. Negative viewpoints like that only serve to discourage others. How would you like it if everytime you fixed or corrected a problem that you’ve caused if people took the attitude of “about time, you idiot!”. You wouldn’t like it very much would you? So don’t treat others that way.
You’re intentionally missing the point. Perfection is a red herring, and a straw man. NO ONE is asking for perfection. The point is that sloppyness is a A-T-T-I-T-U-D-E. That can be carried over to a lot of things. The fact that it persisted over several versions, re-enforces this. The software problem needed to be fixed, but more importantly the underlying attitude needs to be fixed. Just as security is a process, QA is likewise a process, as well as an attitude. You want “2005 to be the year of Linux”? Then you all are going to have to get tougher hides, than present. Because as it stands now, you all can’t take criticism worth a damn.
Well this is opensource, if you want the problem and the attitude fixed, fix it by taking it on yourself. Don’t sit on your ass demanding that others do something while doing nothing yourself. Software development with limited resources (its not like there is 50 people sitting around drumming on the table waiting for someone to report an issue they can fix) will not be able to fix all issues immidiatly. The developers you are attacking for not fixing this particular issue where probably either fixing something else they felt where even more critical to their users.
“Well this is opensource, if you want the problem and the attitude fixed, fix it by taking it on yourself.”
Here’s my problem with this attitude. The OSS community has created expectations through it’s “advocacy”, both pro and con. That includes the development process. e.g. “thousand eyes, all bugs are shallow”. The OSS community really has little room for complaining about being called to live up to the expectations they’ve created. It’s their responsability to getting accepted by the world at large. Not the world at large bending backwards to accept the OSS community. Continue to believe that it’s others responsibility to correct your flaws, and OSS will go nowere.
It just came out this year and there are so many things it can be applied to, that the _FOSS_ world will see dividends from this _Sun_ technology for years to come.
If Sun are smart, they’ll have a whole team of engineers doing nothing but DTrace-ing their whole software stack. With three years of this, Windows will look like a 20-year-old Chrysler K car putting down the highway at 45MPH, while the UNIX world cruises around in Lexus sedans.
So, who here goes to work and comes back at about 5:00, turns on the computer, then turns it off at about 10:00? My computer is on 5 hours out of every 24 hour day, and you’re saying that I should leave it on for the rest of that 19 hours? Obviously someone doesn’t go to work every day.
Yes, a light bulb will go when you turn it on, but how frigging often do we turn a light bulb on/off in a day? The bathroom light will go on and off perhaps dozens of times a day. My computer will not. To suggest that I am at higher risk of damaging my computer because I let it rest for 19 hours of the day is foolish.
Keep dreaming, most people turn their computers off because they are smart and would rather not waste electricity for no reason other than “they’re designed to stay on all the time”.
“To suggest that I am at higher risk of damaging my computer because I let it rest for 19 hours of the day is foolish.”
I’ve had a couple of hard drives fail over the years due to a power cycle. Sometimes they spin down and never see another rotation in their lives. Granted, this is rare, but it does happen, especially with age.
Also, super-cheap computers sometimes come with fans that don’t spin up right away at start. I’ve had fans that need a little push to get them started. I’d bet a lot of PCs out there have frozen fans, and their users don’t know any better.
I guess it isn’t the electronic components that mind power cycling but the mechanical ones. This is where the ‘designed to be on all the time’ idea probably comes from.
Now, with passive cooling and a flash drive…
It’s interesting that the biggest Gnome bounties are from Google.
“No, I want you to stop needlessly ranting. Someone look at a problem, finds it to get a better analysis, and goes to fix it, and all you can do is view it as a failure instead of an opportunity for success. Negative viewpoints like that only serve to discourage others. How would you like it if everytime you fixed or corrected a problem that you’ve caused if people took the attitude of “about time, you idiot!”. You wouldn’t like it very much would you? So don’t treat others that way.”
Somehow I find it extremely hard to believe that this would be your opinion if we were talking a product produced by Microsoft… I think that the point the other guy is trying to make?
Anyone who mentions a problem with Linux, be it large or small, is instantly dismissed as a troll/Windows-lover/loser who’s just “ranting”. Yet if the same comments were being made about Windows, it’d be a whole other kettle of fish.
You know, when Linux still has 5% of the desktop market in 2010, maybe SOME people will step back from it all and think, “Hang on a minute, maybe we should’ve listened to those comments after all.”
But no, go on making things bigger, slower, more complex and overly abstracted. Go on ignoring the larger issues and putting your fingers in your ears when someone has a valid criticism. It’s your loss…
“It just came out this year and there are so many things it can be applied to, that the _FOSS_ world will see dividends from this _Sun_ technology for years to come.”
1) Sun is part of the FOSS world.
2) DTrace is FOSS (CDDL).
I agree it is awesome what Brian made and offered, though.
The point, really, is not that “humans make mistakes.” I can tell you that most nights are dark. That is no excuse, that is not explaining anything, that is not contributing anything to the discussion. That is just arguing for argument’s sake. It is childishly assuming that pointing out a real problem amounts to “demanding perfection” and “not being nice.”
Real coders know their code. They know their skills, their strengths, their weaknesses. They can take a fact, accept it and agree with it. Or counter it with some real argument. For example: “That is a good design, because…”
The fact is, the 300kb XML file is an embarrassment. And there are many like it. This is about an utter lack of design, planning and forethought. There is a science about software design, and there is an immense amount of material on how to design software. So, this is not about “my opinion vs. your opinion” or “just go and do better yourself” or “you are not being nice.”
The Gnome project either gives out free crap, free eternal betas, free amazingly good software, or free something between these extremes. There is no difference to a paid product. Everyone is free to say what he thinks about the software, and does a service to all others if he/she is good at that.
So, this is not a competition of respecting the coders or being nice. This is a competition of making an accurate and unbiased assessment. binarycrusader has made no such assessment. It seems that he is only trying to sabotage others from trying.
I think that many OSS coders are scared of planning anything, because they have no way to know if they can ever succeed in their plans. So, they prefer to add something here, something there, refactor this and extend that a bit. The end result is indeed a mess. And it is used for the same reason as our beloved MS Windows: lack of choice.
I like Solaris. Even if it would take long before I actually use it. And that is only because it is very much designed. The last time before that I saw an OS that had a design was when I met BeOS.
The point, really, is not that “humans make mistakes.” I can tell you that most nights are dark. That is no excuse, that is not explaining anything, that is not contributing anything to the discussion. That is just arguing for argument’s sake. It is childishly assuming that pointing out a real problem amounts to “demanding perfection” and “not being nice.”
It would be different if someone was pointing out a problem instead of insinuating that someone was “sloppy” just because they made a mistake. Since you admit people are always capable of making mistakes, than you must also agree that even careful cautious people can make mistakes. Therefore, the proper attitue is to say, “I hope that future development processes allow the X project to resolve issues like these faster for the benefit of all”. Instead of being a total asshole about it.
There is a science about software design, and there is an immense amount of material on how to design software. So, this is not about “my opinion vs. your opinion” or “just go and do better yourself” or “you are not being nice.”
More insinuation and assumption. Robert Burns reminded us the best-laid plans of mice and men often go awry. Your *ass*umption that just because a mistake was made or that a design problem eventually became visible that no planning was done is wrong. If you had been following the GNOME community at all you would see a great deal of planning and design has gone into the system. Like all designs, they have the potential to fail.
So, this is not a competition of respecting the coders or being nice. This is a competition of making an accurate and unbiased assessment. binarycrusader has made no such assessment. It seems that he is only trying to sabotage others from trying.
I can’t stop others from trying, but I least I can point out the obvious bias and *ass*umptions that others like you are making.
Ximian, RedHat, SUN, and the GNOME foundation are all involved in design, planning, and implementation. Your *ass*umption that none of them are doing any planning due to obvious “perceived flaws” according to you is a gross and outright lie.
“just because they made a mistake”
It’s not a “mistake”. We’re not talking about mistakes. A mistake would be a bug that caused something to crash, or similar — and these things can and do get fixed.
When your desktop is opening and parsing a 300K XML file just to set the volume, that isn’t a single mistake. It’s a reflection on increasing sloppiness, lack of attention to elegant design and following buzzwords rather than doing things the cleanest way.
If it was JUST that, yes, this would be a whine. But there are MANY things like that in GNOME, and they’re not going away. This is just one example, symptomatic of a wider problem.
Of course some design has gone into GNOME. Of course there are some good aspects. But there are a lot of problems, a lot of hacks and areas which haven’t had any elegant design or forethought.
The Linux desktop now is huge, complicated, abstracted among too many layers, full of obscurities and remnants of the past, getting increasingly bloated and difficult to work out under-the-hood. It’s basically becoming another Windows with all the complexity and issues, albeit better and providing more freedom.
But you can’t succeed by just making something that’s, oh, a little bit better. You need something that’s really striking — something that provides many genuine reasons to switch.
And as I talk to many newcomers who’ve tried Linux, and hear their concerns about the sluggishness, the bloat and the vast complexity, I see where it’s going wrong. So if you don’t want to think about the problems, that’s up to you, but I’d suggest you spend time with some newcomers to Linux who’re NOT happy and start to see why, in 2005, we’re still at a microscopic share of the desktop market.
This is what I mean by sloppy design, and this is just not understandable, or bearable. This is ugly. This is a reason to be angry and disillusioned:
http://live.gnome.org/MemoryReduction
There have been machines with a GUI, and those machines had 1MB RAM for the whole system. Still one could easily draw, write and do small-scale DTP with them. Now we have a toolkit that carelessly wastes many times that amount.
There were the times when one didn’t have the luxury of an attitude that says “make it now, make it big, make it far and wide. Make it good later when you can’t make it any bigger.”
This is what happens when nobody stops to think, and this is what happens when nobody cares. So, will they build GTK3 and Gnome 3.0 from the ground up, and maybe this time keep in mind that fatter is not always prettier?
I have an idea. Convert the whole Gnome to the upcoming FLTK 2.0. It can’t be more silly than the present situation. And as the name says, FLTK is a fast and light toolkit. The name promises that the coders have cared, and that they have stopped to think about speed and memory footprint.
They even admit on the page that Gnome is now worse than Windows XP.