Adam Tauno WIlliams wrote:
On Wed, 2009-07-15 at 11:00 -0400, Dan Goodman wrote:
Point 1. This is why Beagle (or Nepomuk, or any "breakthrough" new technology should NOT be automatically turned on at install. However, it should probably install and be easy to toggle on (or back off, if it is found not to work for a user). If beagle was truly totally transparent, that would be different. But it is not -- it uses resources at times that conflict with the user's needs in some cases.
Not enabling new features results in a horrible user experience; note that neither Apple or Microsoft adopt this approach. New features should absolutely be enabled - otherwise the interloper thinks your product is backward crap, it is completely unreasonable to make users run around turning things on.
This is the crux of the problem. Some developers think that they must provide "the latest and greatest" for everyone, and that the user shouldn't have to elect whether or not to use it.
And Beagle, at least, is hardly a "breakthrough" feature as it has been around for something like five years and many releases.
If you want to put 11.x on a NEC Multispeed laptop with dual 720K floppies as its only persistent storage (OK, an exaggeration, but if you want to put it on a fairly old, barebones system...) then you should be willing to remove packages you don't need. Once you are in "the long tail" you should expect to have to do a few things special to get into the game.
Ditto. I'm an advocate of the rule "older than 5 years = onus on you".
A distro should be able to support both legacy HW and state of the art HW, and should be able to be easily installed with a selection option for each style.
But why not offer an install config document, plus the kind of package selection that is available for SLES, so that the issue of installing
Thew is a Wiki; all someone has to do is go create it! I don't believe there would be a single objection to a trimming-your-openSUSE page.
Unfortunately, I lack both the breadth of experience and the time to do so. And apparently so does everyone else, or there would already be one. The trimming should be done by those who are building the distro, with an option for all new features to be on or off at user's choice. And they should be easy to toggle at any time so that people can try them and get comfortable with them before being committed to them.
Point 2. No new technology (that is not truly transparent to the end user) should ever install itself in a way that it is automatically turned on.
Disagree; this equals terrible-usability-experience.
So why don't we default every filesystem to use ACL's? And turning on all new features and giving the user no choice often results in a terrible-usability-experience as well. What you are talking about is probably more of a harder-to-support and harder-to-propagate-your-new-features experience. But the end-user experience should be one where the user is in control, not one that blindly emulates Microsoft and Apple's approach. Microsoft Bob was a new feature that was turned on by default. So was the annoying paper-clip. And each required a Vulcan handshake to disable. And each alienated a portion of the user base, and annoyed others even though they remained. Personally, the first time I switched over to OpenOffice was right after the paper-clip fiasco.
The burden should not be on the user to discover and decide what to do about that new technology.
Most users don't use machines that way; they use what is presented and readily available. Use this approach and they will promptly install a distro that turns all the stuff on. Making life difficult for the normal user (hardware < 5 years old) for the sake of the guy with dual-720k-floppy drives makes no sense.
And developers are clever enough to come up with new features, but aren't clever enough to make them easy to turn on and off, so that both classes of users can be accommodated?
Point 3. Re: supposed beagle slowness. I can confirm that it was slow on a Thinkpad T61 at 10.3 even after initial indexing was done. However, as someone else pointed out, the issue is now reduced to severe slowness only when doing initial indexing, especially if certain excludes are configured. BUT, those excludes, if they improve the average user's experience, should be automatically set at install time, at least as an clearly defined option.
I believe in either 10.3 or 11.0 (at least) it was mentioned right in the installer release notes how to disable it. They pop-up in your face when installing.
That was addressed (rebutted?) in another post. I'll leave it to that post to address this.
Had beagle (and the hotly debated KDE4) been introduced on a buy-in basis, instead of being pushed out to meet the needs and/or desires of the developers, they might not have achieved as much acceptance as quickly as they have. But on the other hand, there would be a lot less frustrated users, and probably more people who gave them a real try on their schedule.
There are some frustrated users, I've seen no evidence there are legions of them. I'm no frustrated and clearly other people who have replied to this thread are not frustrated either. I do not believe the majority of openSUSE users are frustrated.
How would you know? Your decisions about how to handle new and resource intensive features is to base your decisions on what you believe. But many users don't complain, they just switch. Or never start in the first place, and go elsewhere. And when users do express disgruntlement, the standard answer is "you need newer HW and we are taking care of those who want the best, which we are."
The user should have the first, last, and all final says in what runs on their machine, and why. Period, no exceptions.
They already do.
When I installed multiple desktops with my upgrade from 10.3 to 11.1, KDE4 assumed that it was the one I intended to run. But what I intended to do was install several, and try each of them. And I finally had to do a clean install, without KDE, to get my system to where I didn't have constant conflicts between KDE4, KDE3, Gnome and a couple of lesser-known options. And I could clearly see that the footprints were those of KDE4, not the result of something else hosing things up. KDE4 simply assumed that it was the one you wanted, similar to the way browsers used to make themselves the default if you installed them. Notice that they do not do so any longer, (at least FF and IE, for example), because that is not what users wanted. They wanted to be able to try them without permanently changing their system, and without having to read a bunch of tech docs and do a bunch of reconfiguration, to get back to where they started, if they reject the new option. Beagle was similar, except for the fact that it didn't break anything else...it just assumed that I wanted it. And then it slowed everything down on a machine with a dualcore CPU and 1Gb of ram. Noticeably slowed it...and not just in initial indexing. Not until it had been thoroughly investigated and reconfigured could I get it to stop annoying me when I was doing something else. And that only at 11.1...never would stay out of my way with 10.3. System monitor would show it kicking off and indexing every site I visited, etc., slowing the system to a crawl. That was a horrible user experience.
And please, don't throw out bug fixes as a counter-example. I am talking about packages only, and then only those that introduce new techniques at the expense of configuration and resource demands. They are what I am talking about when I say that they should be installable (as an option, preferably), but should NEVER be turned on automatically just because they are installed.
Again, installing without enabling results in a horrible [desktop] usability experience. "I installed it, now why isn't it working?"
Not if you clearly give the user an option to install and activate, and another to install in standby mode -- easy to turn on, but not automatically enabled. And not if you make it really obvious how to turn it off. Chkconfig/inssrv would be fine, along with an easy to find preference item.
Right now, Suse is my distro of choice, both because of personal preference and professional needs. But if it, or any distro, got to where it was dictating to me what my desktop "should" look like, or how my data is organized, I would drop it like a hot rock and find a distro that treats new things as options to be chosen by the user,
There is always Gentoo! For those who *LOVE* a truly horrible user experience.
Yes, and if you don't want to buy a gas-guzzler from me (Honest John's Used Cars), you should change your mind when I point out that if you bought a Hummer, your gas experience would be truly horrible. Sorry, but giving an example of something that is worse does not address the issues around this situation, and does not address that better alternatives may exist. This is a smoke-and-mirrors argument, as Faulkner (and I think Shakespeare) said: "full of sound and fury, signifying nothing...".
Beagle is a solution to a problem for SOME, not ALL. That is why it should not be turned on automatically.
X is a solution to a problem for some, not all. Thus it should not be enabled by default. Same with Avahi, Hot Plug, ACPI/APM, the screensaver, and DHCP. Why should I have to be bothered to disable the screensaver?
X is not enabled by default, and locked in. It is the only easy way to do a GUI on this distro, but there is an easy to obtain text mode (runlevel 3) as well. Most users want a gui, and turning it off is straightforward and documented everywhere - almost.
Sustainable progress cannot be obtained at the expense of costing users time and effort that they have not signed up for. People contribute in various ways to open source based on "enlightened altruism", not unbridled, totally self-less altruism.
Nope, "Sustainable progress" is created by offer a fantastic state-of-the-art feature-on-par-with-OS/X-and-Vista desktop. Which thankfully openSUSE does RIGHT NOW! It is an awesome distro.
Sustainable progress for me, and for many, is not defined by "keeping up with the Joneses". It is defined by having robust and easy to use functionality that is less robust, more expensive, less functional and harder to configure everywhere else. If you want to offer me every feature on Vista and OS-X, fine. But don't activate them and assume that I will want them, or won't mind having to figure out how to turn them off...after all, I must be prepared to suffer for not playing the HW arms race like M$ and Apple...according to you. I reject that idea totally...and I don't know many people who would choose this, or any distro, because it has Vista Aero-like appearance. But you sound like a marketing guy...there are no problems, only opportunities. This stuff is GREAT! Only the Luddites can't see the obvious advantages. Etc., etc.
As a dev, you may be smarter than the average user. But the only way you are going to win a base of users is to educate and convince them, not by corralling them and then hoping that they stick around once you have them running your package.
Devs, you've won me by providing a solid and current product. Keep up the great work.
Rah-rah. Hip-hip Hooray! Some have and some haven't. And I still say that devs do not have the right to assume that a user should be locked into a feature, just because they want to try it. How many people would have complained about having to turn on the paperclip in Office, if it was made available but not mandatory? And why did M$ FINALLY provide an option that was easy to find from the clip itself, for turning it off? Then I don't mind it being on by default. But when you had to hunt for documentation telling you how to turn it off, and then hunt for the menu options, that was not acceptable, and produced a user backlash.
I really don't want to leave this distro,
I see no reason to leave and very much intend on staying.
I see reasons to stay, and reasons to leave. And one of the reasons for leaving would be if the packagers persisted in the philosophy of turning thing on, eating up resources, and making them difficult to turn off. And so would many others, or they would just become Ubuntu users in the first place. I haven't heard of any Ubuntu issues where systems got bogged down by new features implemented the way Beagle and KDE4 first were. Fortunately, the decision on Nepomuk shows that someone is listening. But you must feel that not turning on Nepomuk by default in 11.2 is a big step backward for the distro, based on what you have said elsewhere in this discussion. Why aren't you railing that the distro is in decline because it will distribute a sisabled important new feature (installed, but not turned on, as I understand)? You give me the impression that you dismiss any criticism of things done other than the way you want (automatically implementing all new features) as simply being a disgruntled minority, or people with old HW, or people who stand in the way of the enlightened new users who want, no need, to have all of these new features in order to choose this distro over Vista. If I am wrong about that, please forgive me. But all I hear from you is that any new feature a dev has developed that is good enough to be offered is good enough to install by default, and that turning it off is an anomaly that can be handled by offering after the fact documentation on how to turn it off -- documentation that requires many steps, many of those beyond the capability of newer users. So how would you ever be convinced that there are enough people who don't want all new features fully activated? Could you ever be convinced if enough people spoke up? Or are you thoroughly convinced that no "right thinking" user would ever refuse any of these new features, once they are activated and after they have been thoroughly wrung out by the user community? Dan Notice: This communication, including attachments, may contain confidential or proprietary information to be conveyed solely for the intended recipient(s). If you are not the intended recipient, or if you otherwise received this message in error, please notify the sender immediately by return e-mail and promptly delete this e-mail, including attachments, without reading or saving them in any manner. The unauthorized use, dissemination, distribution, or reproduction of this e-mail, including attachments, is strictly prohibited and may be unlawful. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org