Home > macOS > Report on Tiger WWDC Build with Many Screenshots Report on Tiger WWDC Build with Many Screenshots Guest post by Alexandros Roussos 2004-07-01 macOS 41 Comments On this report I will detail some Mac OS X Tiger WWDC build (8A162) new features that will be illustrated with many screenshots. I will try to focus on features other sites did not report yet. 41 Comments 2004-07-01 12:09 am Anonymous i knew that if there was a place to find this, it’d be on os news 2004-07-01 12:55 am Anonymous why has no-one i read about refered to the speech features 0f OS X ? http://www.apple.com/macosx/tiger/voiceover.html sure it isn’t new, but intergrated into the OS is a big deal no? 2004-07-01 12:56 am Anonymous >why has no-one i read about refered to the speech features 0f OS X ? Because Steve Jobs did not mention it at all in the keynote, so it got undercoveraged. 2004-07-01 12:56 am Anonymous I remember that people were up in arms when the previews of KDE 3.3 showed a “search” function in the control center. “So complicated that it needs a search!” they said. Now, Apple put’s one in Tiger: http://mirror1.macosxrumors.com/images/040630/prefs_search.jpg Of course, nobody makes a peep. 2004-07-01 1:06 am Anonymous Funny how that happens, an idea is crap unless it is on your OS, then it’s cool all of a sudden. I personally never thought KDE doing it was a bad idea, really… I personally think it’s a great idea. I can never find where to set up some networking things on OS X at times, Linux or XP. I wish they all had this feature, esp. XP 😉 2004-07-01 1:07 am Anonymous @Rayiner Hashem I remember that people were up in arms when the previews of KDE 3.3 showed a “search” function in the control center. “So complicated that it needs a search!” they said. Now, Apple put’s one in Tiger: … Of course, nobody makes a peep. Yes, because unlike KDE, Apple doesn’t have 30+ pages of options. The sad fact of the matter is that KDE’s prefs system *NEEDS* a search engine. While Apple’s really doesn’t *need* one. It is an added *benefit* but not a necessity. That’s why people aren’t saying much. In addition, the way Apple is doing it seems more intuitive and a lot more visually appealing. 2004-07-01 1:10 am Anonymous To a degree i think search is over-rated. someone on the gnome project refered to a system where the user should rarely need to open the hard-disk because applications save data in a very anal manner so every app knows where the data it can use is. BUT that search thing could well make supporting a mac user even easier. AND even an organised person needs search once in a while and, by god, when you do i think you appriciate a good system with ready-made indexes. If only creating a file automatically updated the database locate uses it is not much effort but i don’t want to mannually update it! Those virtual folders in mail are good too, i love them in evolution. the only thing a feature really needs to avoid me hating it is the option to turn it off or remove it if i don’t like it. Once we have that .. bring on as many new features as you like! I love that about firefox, going through the extentions and taking what i like and dumping what i hate. 2004-07-01 1:21 am Anonymous Tiger’s preferences panel has 25 icons, which means at least 25 pages of options. KDE’s has quite a bit more, but it’s certainly not an order of magnitude difference. In any case, if the preferences panel becomes a one-stop shop for control settings, it’s bound to get large, and it’s not surprising that it is large, even in MacOS. Any way you cut it, a search-based interface to the control panel is a *good* idea, unless you have a trivially small number of options (which no modern OS since BeOS has had). As for more visually appealing, I’ll buy that, but more intuitive? http://www.avenheim.online.fr/kcontrol3/search.png The only difference between the two is the placement of the search bar (above vs below). And I’m docking points from Tiger for not taking you directly to the relevent option, but instead just highlighting the general category. 2004-07-01 1:25 am Anonymous You are missing that when searching in the OS X Pref Panel, the results are also displaying for ALL help files which may contain references to the search in relation to a Pref Panel selection. So you could easily do a search that could apply to many different panels. It should DEFINITELY not auto jump to the one you think is most appropriate. The context for a searh is much larger then the names of the panels and the few words which appear on each of these panels. 2004-07-01 1:32 am Anonymous To a degree i think search is over-rated. Having to manually organize your files on a computer is ridiculous at face value. I want to be able to absent-mindedly throw stuff at my computer, and let it organize things for me. I can’t do that with my desk, because I don’t have a personal maid, but I should be able to do that with my computer. 2004-07-01 1:35 am Anonymous Never mind. I was ignoring the little “search results” popup next to the search box, and just looking at what was highlighted. So in that case, the functioning of the two systems is identical. How, then, does OS X’s “seem more intuitive”? 2004-07-01 1:51 am Anonymous BINGO!!!!! computers SHOULD organize your files for you. a data-centric system that organizes based on what YOU want automatically in a manor that is meaningful to your work flow is what computers should have been able to do for years, yet it has taken MS and Apple all these years to finally get that to the customer. 2004-07-01 1:57 am Anonymous I don’t need search, my files are organized, I can find things faster by navigating to them…. This is actually a pretty good point. I hope with all the new search technologies being researched people don’t underestimate just how useful it is to keep files in a separate folders. This is a lesson I suspect a few are going to learn the hard way. This is one of the reasons I think linux file layout (FHS) needs to be completely redesigned. 2004-07-01 2:07 am Anonymous How does this work? Surely, for example, a computer can’t tell that a photo I take is of Uncle Bob in Amsterdam. Don’t I still have to manually enter information about the files, so that the search parameters become useful to me? Same with text/word processing files. I know you will be able to search text, but that’s no different from the situation today. If they are going to advance this, doesn’t someone have to pick out what a particular text file is about, and contextual information about it, for example other related documents/emails/images/sound. Seems that this process is just as time consuming (probably more so), than organising files in a heirarchy. And given that all the work is done up front, it is likely no one will bother, and will become useless. Am I missing some computer magic that’ll do all the categorising of files/entering metadata for you? 2004-07-01 2:11 am Anonymous Redesigning the UNIX FHS wouldn’t make any sense. The FHS is really optimized for the OS, rather than the user. That’s OK, because it’s always been a key idea in UNIX that the user’s files (which are in /home/username) are seperate from the system’s files. Even MacOS and Windows, which used to give the user full access to the filesystem, have now adopted the UNIX approach of mostly focusing them on their individual user directories (Home and My Documents). The simple fact is that the hierarchical filesystem is great for the OS, but confusing for the user. Hierarchical structures in general are confusing for the user. Ever wonder why iTunes presents the song database as a single flat listing? That’s why search-based UIs make a lot of sense. They use concepts that people used everyday: categorized sets, instead of concepts that people don’t use very often, hierarchies. 2004-07-01 2:21 am Anonymous well, in the future you will have the meta data as part of the save dialogue. then you will have the location of the file hidden from you and you will have to add meta data to it that will designate a “pile” or “stack” or “smart folder” for it to appear in. that is much faster than navigating to a specific folder to save the file in, especially in a more complex organizational system. these folder types which are really saved Queries will also be able to have meta Queries in them so that you can further organize the data. 2004-07-01 2:28 am Anonymous um, maybe someone forgot to tell the dashboard dev team, but F12 is eject on apple laptops. does this mean every time i want to write a note on a hardware rendered, gui-licious sticky i have to eject my felix da housecat CD? 2004-07-01 2:32 am Anonymous umm, reassign the frigen hot key then!!! 2004-07-01 3:03 am Anonymous >umm, reassign the frigen hot key then!!! Sorry, but this is not right. The OS manufacturer is *responsible* for an interface that makes sense for the features that ship with the OS. If by pressing F12 does TWO different things by default, then there is a problem in the defaults. Remember, “defaults matter”. 2004-07-01 3:13 am Anonymous “The FHS is really optimized for the OS” then “Hierarchical structures in general are confusing for the user.” They are about as confusing for a user as using a file cabinet. Pictures in one cabinet, documents in the other. So if a user is looking for vacation photos, they go to $HOME/pictures/vacation or where ever. The people that can’t grasp the concept of a _folder_ are people that use a computer less once a week to get on the “innernet”. If I wanted a system designed for the lowest common denominator, I would go back to high school. I am in favor of keeping a hierarchical file system, I just don’t agree with the locatsions used in Linux. eg. /usr/bin contains 1100 files. expect, which, wc, diff, grep etc. are in the same directory as everything KDE traceroute is /usr/sbin, but ping is /bin/ping? /bin/gzip and /bin/gunzip but /usr/bin/zip and /usr/bin/bzip2? The command line nature of Linux is all the more reason to have more intuitive locations for things. How hard can it be to separate GUI aps (gimp, gaim etc.) from command line utils like telnet and tail? it is not like changing this would require a rewrite of the kernel or even break compatibility with anything. The file layout simply was not designed for what it is being used for today, back then it was common to have stand alone binaries like emacs and ping designed for shell use, today many GUI applications like OO.o and Firefox need folders. If I have a few mp3 and a couple vids, it is fine to just throw them in a /media folder, but for thousands of pictures, vids, songs of different things I have to create more folders to better organize what is there, this is what FHS (Linux) is not doing. The majority of my complaints would probably be solved with an addition of an /apps folder to FHS. It also wouldn’t break anything. 2004-07-01 3:18 am Anonymous I don’t think it’s a problem. On my Macs, you have to press and hold F12 to eject the CD. Just a press doesn’t do anything. So I think you can simply press F12 to access Dashboard and press & hold to eject. I’m quite sure the Dashboard development team was fully aware of what pressing and holding F12 does. 2004-07-01 3:30 am Anonymous @Rayiner hashem “Never mind. I was ignoring the little “search results” popup next to the search box, and just looking at what was highlighted. So in that case, the functioning of the two systems is identical. How, then, does OS X’s “seem more intuitive”?” Because if you type, wlan, 802.11 o anything similar it will take you to the airport preference panel.. If you type wallpaper to the desktop & screensaver panel… I think for a newbie/switcher is more intuitive… @Dumbkiwi “How does this work? Surely, for example, a computer can’t tell that a photo I take is of Uncle Bob in Amsterdam. Don’t I still have to manually enter information about the files, so that the search parameters become useful to me? Same with text/word processing files. I know you will be able to search text, but that’s no different from the situation today. If they are going to advance this, doesn’t someone have to pick out what a particular text file is about, and contextual information about it, for example other related documents/emails/images/sound. ” Actually, if you have seen the keynote, you should know that Spotlight find inside the content of some filetypes,emails, text, word, html, pdf and a long etc… It would be too difficult to make a search engine that has realtime face recognition & indexation technology to find the correct image of your uncle bob .. But files with text content inside can already be searched in tiger.. Was demoed in keynote.. 2004-07-01 3:46 am Anonymous They are about as confusing for a user as using a file cabinet. Have you *seen* the average person’s desk? There is a reason people pay secretaries to organize filing cabinets. /usr/bin contains 1100 files. expect, which, wc, diff, grep etc. are in the same directory as everything KDE Because they are all binaries. Why does the user care where the binaries are, as long as the shortcuts in their start menu work. Keeping everything together makes things easier for the system, because it simplifies the $PATH, etc. traceroute is /usr/sbin, but ping is /bin/ping? /usr often on a separate partition, so only the absolute minimum should be kept in /bin. If ping wasn’t potentially useful in debugging a problem with mounting /usr from an NFS share, it’d be in /usr too. /bin/gzip and /bin/gunzip but /usr/bin/zip and /usr/bin/bzip2? Again, only the minimum necessary functionality should be kept in /bin. Gzip is the de-facto compression format in Linux, so it gets to be in /bin. Less-often used stuff is relegated to /usr. The command line nature of Linux is all the more reason to have more intuitive locations for things. Even the UNIX CLI is designed with the purpose of abstracting the physical location of system files. That’s why there is a $PATH, and why ‘man’ doesn’t take a man *file* as a parameter. And the UNIX FHS is very intuitive if you understand that it was designed for a networked environment. How hard can it be to separate GUI aps (gimp, gaim etc.) from command line utils like telnet and tail? What’d be the point of doing this? Some misplaced sense of cleanliness? If I have a few mp3 and a couple vids, it is fine to just throw them in a /media folder, but for thousands of pictures, vids, songs of different things I have to create more folders to better organize what is there, this is what FHS (Linux) is not doing. The system doesn’t need the files to be organized. From it’s point of view, a directory with 1000 binaries is as easy as a structured tree. And a raw directory tree isn’t a good abstraction for the user anyway. The search-based view is quite natural to UNIX. Think about it: what is APT? A search-based interface to package files! What is man? A search-based interface to man files? What is $PATH? A (primitive) search-based interface to binary files! What’s the point of creating some folder hierarchy when the system doesn’t need it, and the user doesn’t need it? 2004-07-01 4:26 am Anonymous A hierarchical filesystem is clearly too complicated. Why has everybody forgotten the forth blocks system? Its a much simpler and more intuitive way of storing data. 2004-07-01 4:32 am Anonymous The Powerbooks have their own eject key. 2004-07-01 4:42 am Anonymous The command line nature of Linux is all the more reason to have more intuitive locations for things. How hard can it be to separate GUI aps (gimp, gaim etc.) from command line utils like telnet and tail? it is not like changing this would require a rewrite of the kernel or even break compatibility with anything. One of the reasons why FreeBSD feels more “whole” is that it does precisely that, and the is organized consistently and thoughtfully. 2004-07-01 5:13 am Anonymous A flat database oriented design makes sense for files that have a strong categorial properties like music and video, but it doesn’t make much sense for much else. I think these concepts should be used for their proper problem domains and not attempt to apply them to everything. Don’t try and hammer a square peg through a round hole. Newton OS already has a 100% database oriented file store (this was way back in the 90’s mind you). One of the first extentions people wanted for this thing was to allow heirarchical filing inside of categories. 2004-07-01 5:48 am Anonymous i’m typing this from a PowerBook 1GHz DVI right now and eject key = F12. to get the actual F* functionality you have to press the [fn] key, but this hasn’t been the case for Expose, which works whether or not the [fn] key is pressed (F9, F10, and F11 do not have any function other then use as F* keys). all revisions of the iBook since May 01, and all the powerbooks from 01 until 02 (not the aluminum ones) have the F12 key as the eject button ( http://developer.apple.com/documentation/Hardware/hardware2.html ). all the new aluminum revision powerbooks have a separate eject key. i don’t like the press or press or hold option. that seems like a cheap way out. i hate to say just make it F8 because that will get all the konfabulator zealots screaming, but i say just make it F8. but that’s just my $0.02 2004-07-01 6:00 am Anonymous dude, relax… it’s just a beta if u have such an issue with it, file a bug report here: http://www.apple.com/macosx/feedback/ 2004-07-01 6:02 am Anonymous Because they are all binaries. Why does the user care where the binaries are, as long as the shortcuts in their start menu work. Keeping everything together makes things easier for the system, because it simplifies the $PATH, etc. Which is it? If it is in the menu and the user does not ever have to know where it is, then why must it be in $PATH? And would it be that hard to add one more directory to $PATH? Here is the $PATH on my box /usr/bin:/bin:/usr/bin:/usr/local/bin:/usr/X11R6/bin/:/usr/games:/usr/ sbin:/sbin:/usr/local/sbin If they do need to be in $PATH, it is not like adding /apps/bin is going to run me out of swap space. Create the folder in /apps/MyApp and throw the binary or a static link in /apps/bin. Also leave /etc for OS stuff and services, /etc/gimp /etc/gnucash have no place in /etc, fstab, hosts.allow, xinetd.conf, and resolv.conf do. How would like to manage Trillian from C:windowssystem32etc rillian.conf ? I still don’t see any valid reason why Linux cannot separate GUI apps from console apps. You might say “users” don’t need to know where stuff is located on the disk but I don’t agree, how many people reading this do you supposed have manually “explored” “Program Files”? I am willing to be just about all of them for one reason or another. First, I try to avoid the menu, it is more time consuming than desktop icons. Second, not everything I install creates a menu entry. I often install an RPM and then not even know the name of the binary that I need to run it. Su root; update db; locate blah ; no dice So I consult the RPM DB to see where the package was installed, simple enough. #rpm Usage: rpm [-aKfgpqV] [-aKfgpqVcdils] [-aKfgpqVcdilsaKfgpqV] [-aKfgpqVcdilsaKfgpqV] [-aKfgpqVcdilsaKfgpqVK] [-aKfgpqVcdilsaKfgpqVK] [-aKfgpqVcdilsaKfgpqVKi] [-aKfgpqVcdilsaKfgpqVKiv] [-aKfgpqVcdilsaKfgpqVKiv] [-aKfgpqVcdilsaKfgpqVKiv?] [-a|–all] [-f|–file] [-g|–group] wow, OK, so I write down the command for the next time I need it. Example#rpm -qil gimp (which BTW lists almost 900 entries) So I find the installed binary, create a desktop shortcut for it, brows again for wherever it stuck the icon file, and finally I now have a working shortcut to the app. Good thing I didn’t need to know where the installer threw all the files. Because knowing it was going to be installed to /apps/MyApp might have been more than I was ready to handle. FHS is great for services, but for GUI apps it is a train wreck. 2004-07-01 8:26 am Anonymous I intend to buy Tiger (currently a Panther user).. however if the minimum requirement is a mac with Firewire… then I need to get a new Mac too ! Which is a shame, as my 366mhz iBook is fine with Panther, and OSX upgrades normally run faster so I dont see the spec of this machine being a problem for Tiger apart from the lack of a Firewire port. 2004-07-01 8:42 am Anonymous Don’t worry, CdBee… just grab XPostFacto from macsales.com when Tiger is finally released and I bet you’ll be able to get it running on your iBook just fine. I have Panther running beautifully on a Beige G3 AIO which is anything but supported. 2004-07-01 9:37 am Anonymous “It would be too difficult to make a search engine that has realtime face recognition & indexation technology to find the correct image of your uncle bob ..” However, digital camera images are the kind of thing that really is a problem, as the filenames are just numbers. We have always had search engines for finding text files. What about the composer who has thousands of samples of drum sounds? I think a lot of fuss is being made about something that doesn’t solve the real problems. 2004-07-01 3:03 pm Anonymous I wouldn’t change the unix filesysytem for programs if I could do that. I WOULD impose some sort of order on the user though. A fast & easy way to force files of certain types into certain folders would create a rigid structure that limits the scope of the search. This makes it quicker, and it means that a program opens and it knows where ALL the files it can use exists. All programs could even offer a search of this limited area using a standard interface. This standardisation could be a BIG part of a shift towards better searching systems. All programs, in my view, could reasonabley make the requirement of entering a few words of metadata COMPULSORY. These few words could apply to one file or a batch of files being saved onto the harddisk. So you could save pictures of uncle bob in amsterdam, along with all the rest from that holiday, and be forced to enter the keywords “holiday” & “amsterdam”. They could enter more if they wish, but those 2 words would go a long way towards finding ALL of the holiday pictures. As users get more used to this and see the benefits, a number of people would make the effort to enter a few more words. A lot of people I know include one or more obvious discriptive words in the actual filename of the picture. Searching filename for keywords too, would make this system even better. Better search tools AND greater limits on where things can be kept, to attack the problem of lost files from two angles. 2004-07-01 3:15 pm Anonymous >>> Good thing I didn’t need to know where the installer threw all the files. Because knowing it was going to be installed to /apps/MyApp might have been more than I was ready to handle. >>> If you want to know where a program is located, you can just do: whereis PROGRAM Per your example, “whereis gimp” would show you the location of the binary, config file, and man page. If you want just the binary location, you would do “which gimp”. So yes, it is very easy to find where a file is. 2004-07-01 3:59 pm Anonymous These few words could apply to one file or a batch of files being saved onto the harddisk So all pictures (.jpg, .gif, .png) should be stored in a big database and all I have to do is add metadata “bob in amsterdam” before saving them? OK, so instead of naming one file I have now added metadata to 300 pictures. What if I don’t remember the description I added, I can’t browse there, but a good dB client should be able to search out all pictures and list only metadata discriptions. After I find the description I need, I narrow my search criteria to that, and bam, I have what looks like a folder containing pictures of bob’s visit to the Netherlands. How was that easier than /pictures/bob/amsterdam again? Why can’t we just add metadata and ID3 tag support etc. to existing search tools to the existing method. It would do both things. ############### To Anonymous (IP: —.wchryh01.nj.comcast.net whereis is only useful when you already know the name of the binary, if I knew the file name I wouldn’t have to be looking for it. Gimp was only an example, but not all applications use the project name for the binary. A better example (maybe) is net-snmp and net-snmp-utils. #whereis net-snmp nothing, because the binaries that package installs are snmpwalk, snmpget, snmpset, snmptrap etc. That package installs about 17 binaries, many have snmp in the name, but others are mib2c and fixproc. Most of the conf files are in /usr/share/snmp/, and not /etc. 2004-07-01 4:07 pm Anonymous It’s been mentioned that adding metadata entry into the save dialog would be a good idea. I agree, it seems pretty obvious. It would be nice if all Cocoa and Carbon apps were to inherit this functionality. But something that would be nice would be to have the program pre-add or pre-fill some values. For example, if I were working on an Xcode project, it would be nice if my source files were tagged with the project name, and which target(s) they’re used in and so on. And if I were to change some of the project’s properties, Xcode would update the metadata for all the files accordingly. Honestly, I don’t think this is asking too much — if you’re going to do metadata, you might as well do it seriously. It’s also important that simple APIs are added to Carbon and Cocoa to make adding and changing the metadata as simple as working with an NSDictionary. I’ve done some BeOS metadata programming and while it was not super-hard, it was decidedly non-trivial. I think it also matters to be able to make easy batch changes to groups of files, from the finder or from metadata friendly apps. 2004-07-01 4:13 pm Anonymous “I intend to buy Tiger (currently a Panther user).. however if the minimum requirement is a mac with Firewire… then I need to get a new Mac too!” I wouldn’t worry about system requirements as of yet. The OS is still being developed. The system requirements that have been released thus far are for the developer’s preview. The final version may have a different set of requirements. 2004-07-01 5:41 pm Anonymous Having used KDE 3.2 for a year and OSX for a few months I’d have to say that yes, KDE’s control panel _is_ convoluted. I find OSX’s System Preferences to be much easier to understand, and the search will help newbies a bit more. Having a system-wide search function available will be a very good thing for beginners, and indeed everyone. Damien 2004-07-01 6:13 pm Anonymous First poster: I don’t need search, my files are organized, I can find things faster by navigating to them…. Second poster: This is actually a pretty good point. I hope with all the new search technologies being researched people don’t underestimate just how useful it is to keep files in a separate folders. No, this is actually a pretty bad point. “Organizing” your files imposes JUST ONE of arbitrary hierarchies. Like: /music/mp3/Rock/Nirvana What if you want to select and (hear/work on/store/whatever) based on another hierarchy? Hint: you can’t! You have to manually go and select multiple folders and/or files. I now want, say: /music/Rock/Nirvana regardless if its mp3 or album covers, or song lyrics etc. Yes, I could have things stored in this way instead of the other. But 10 minutes later I might need yet another “view” of my files. I cannot reorder actual directories on the fly! If anyone mentions soft links, he is an imbecile. This “solution” is already you doing the work of the computer yourself, and in a kludgy way. Also, you people are thinking of the 60-120GB drives of today. Think also of the 2TB drives in 10 years time. Which brings us to our second point. What’s even worse, is, when manually organizing your files you do repetitive, dull work that the computer is supposed to do for you. 2004-07-01 7:58 pm Anonymous FireWire is probably only required if you are going to import your user from an another computer. They probably just don’t have all the documentation cleaned up yet. As for DVD. You could probably get away doing a remote install of Tiger onto a Mac that doesn’t have a DVD player.