Slashdot Mirror
Major Problems With Safari
kuwan writes "There have been many problems reported with Safari on Apple's discussion boards. The two most prominent are that option-clicking on a link to download can replace your Home folder with the downloaded file, effectively nuking your Home folder. The other has been reported as a printing problem, but is far worse. The printing problem occurs because Safari deletes /tmp, which is a link to /private/tmp."
Man, we all know the software companies have to say that nothing works and everything should be used at our own risk, it's just the lawyers that make them say it, but (nudge nudge wink etc) it actually works like a charm with no problems whatsoever.
When you're warned about all software, even final retail copies, it's hard to remember to take beta warnings seriously.
Remember when you found out your teachers lied to you about all illegal substances turning you into a zombie or worse? You were smart enough to figure out what's OK and what's not, but some people figured they should just ignore all the warnings because they don't mean a thing. That's right, the points don't matter, just like Netscape.
> isn't /tmp a link to /private/tmp anyway
/tmp, which is a link to /private/tmp." The problem IS that the link is deleted. Not hard to fix (if you know how to use the CLI), but annoying.
Exactly, that is what it says above; "Safari deletes
Yeah, too bad we can't mod the story as redundant. The purpose of a beta release is to find bugs, so the Safari beta seems to be doing its job. It's better to have a buggy beta than a buggy final release.
The responsibility for that problem lies squarely on the shoulders of the Mozilla guys. Back when I was young, "beta" meant "feature-complete, but not yet debugged." If a product is in "beta," that meant that it was absolutely not going to get any new features before release.
The Mozilla guys kept glomming features onto their browser for months and years. Eventually they got rid of the term "beta" and started calling them "prereleases" or "milestones" or something, but the fact remains that it's an awful practice.
Apple has a history of treating betas like betas. The Mac OS X public beta didn't get any major new features when it went to 10.0. iSync beta didn't get any major new features when it went to 1.0. And I hope, oh I hope, that Safari doesn't get any major new features before it goes to 1.0.
I write in my journal
For Mac faithful, try Chimera Navigator which is astonsihly based on Gecko yet worlds apart from Netscape. It's a nice example of a rapid development project benefitting from the OS X environment. 0.6 is a major advance.
I use Opera and like it, but you do have to pay $40 for it, and we have three machines. I'm a little worrid about Opera's apparent feature creep.
I don't know what Chimera's future is, but it's free and GPL. I wish Safari all the best but will wait a little. And WHY with Aqua have they still not dropped that awful brushed metal look??? Chimera does a better job of Aqua than Apple's own product.
I've tried to replicate this bug and cannot, when noone is giving specific information that will allow a tester to replicate a bug that usually means that something else has happened. I'm a software developer who been through many alphas and betas and I can smell hysteria here. It's human nature to assume that computer problems are caused by the software that's currently being tested. Sometimes this is true, and in those cases the effect can be reproduced and hopefully fixed. Many other times the software in question cannot be prodded into recreating the error, and it is often something else (like something the user did that is not related to the software being tested).
/private/tmp in /
The second part of this phenomenon is the hysteria -the (possibly erroneously reported) bug looks bad and pretty soon everyone's talking about, it seems like everyones got it and anything that goes wrong with a person's computer is due to the program.
You also get a significant "joiner" effect -people who didn't like the software, or company or whatever begin clamoring about the (now many) debilitating bugs.
I'm not especially impressed with Safari one way or another, but I've been using it for many hours, doing all the things that are supposed to cause problems, and I still:
1) have a home directory
2) have a sym-link to
3) can print
What am I doing wrong!!!
Forgot my UID so I am anonymous...
-Aaron
Well, for one thing, prior to Safari, KHTML (in the guise of Konqueror, I guess, although I've never run it under Mac OS X myself) ran as a KDE application under X11 in OS X; now, Safari runs as a native application in Aqua. I don't know how much work this took, but it obviously differentiates Safari from KHTML. The UI of Safari is noticeably different from prior implementations of KHTML; I don't remember having seen the SnapBack feature in KHTML before Safari, although I could be just out of the loop; and, as one would expect, KHTML prior to Safari did not include the Address Book and Rendezvous integration that Safari now has. These facts, among others, indicate to me that, while Apple might still have a ways to go with Safari (hence the "beta"), they did put at least some significant degree of work into it before releasing it, and I'm not sure there's any evidence, in the absence of inside information, that Safari was a rush job.
With regard to the two major reported bugs, I don't think it ever makes sense to tell otherwise intelligent people who've just had their home directory wiped out after trying to download a file that the bug is nonexistent. In reality, however, these bugs seem to be being reported sporadically at best, and there may be some specific set of circumstances outside of Safari which cause these bugs that we're unaware of. I think the best you can do, then, is acknowledge the usual caveats that go along with using a beta; unless you've actually experienced the bugs, however, I think running scared from Safari might be overdoing it a bit. As always when dealing with software before its official release, caution is the better part of prudence.
I am beginning to wonder if open betas are really the right way for software development to be moving. Recently someone mentioned how easy it is to get beta software these days, and I realized that they were right, it wasn't long ago that beta software was something you had to apply for and hope to be chosen to recieve. I think that this uniqueness prevented less knowledgable users from gaining access to the software, at least without some effort, which meant that if you got it you generally knew what it could do, and were prepared to deal with the consiquences. Perhaps more software should go back to a more closed beta to prevent these sorts of situations from occuring.
Incidentally, did you know that SPAM is concidered a treat in Hawaii? They have some resturants that feature it in $20 a plate dishes, and it usually sells out on paydays. I have heard that it was popular at truck stops as well.
Degaussing scares the bad magnetism out of the monitor and fills it with good karma.
Well. Yeah. But it could use some tabs... :-)
According to some of the developers who wish to remain nameless (no, Hyatt isn't one of us... publicly) Safari will get tabs over our dead bodies. We didn't omit them because we didn't have time. We left them out because they're a terrible UI design. To see this in action, just open six or seven tabs in a Chimera window. You can't even read the titles any more! Tabs are pretty useless at that point.
Read my lips. Safari will never have tabs.
I like Chimera a lot, but I still prefer galeon because Galeon puts close "X" buttons on each tab. If Chimera did that it would be perfect for me.
I usually close tabs from the keyboard anyway, but for the mouse, Cocoa Gestures are much better than tab Xs (not to mention window widgets), at least if you have a wheel mouse.
I hereby place the above post in the public domain.