|
On Nov. 26, 2001, DesktopLinux.com published my impromptu article which suggested that Linux is not quite ready for use by the general public as a Desktop environment. This is my concluding post to the discussion thread that ensued.
Date: December 12, 2001
I am the original author of the article that started this entire thread. Since the dust has settled, I will make this final post. Please bear in mind that my 'article' was actually not written as an article. It was simply a reprint of an impromptu email message I sent to a couple of people. If I were to write a real article on the subject of the Linux Desktop, it would be much more thorough. As for the replies I've seen here, I will respond specifically to sgaap's last post since it is the latest post in chronological order as of the moment I am writing this and since it represents a dispassionate simple cut at a few of the salient issues raised by others in the thread. I will then conclude with some general remarks.
Reponse to sgaap
> Mozilla is stable and pretty fast (it has no preloading in Point well taken. I overemphasized AbiWord. Had there not been so many other issues confronting the newbie's I was trying to help, I would have installed WordPerfect or StarOffice and tried those out. I am not saying yet that I believe they would have solved the situation for the newbies. I *have* tried both of these products, but my experiences with them were a few *years* ago (at least three years ago). So, I won't say yeah or nay until I give them a fresh try. I fully mistrust all the positive remarks people make about those (or any other) programs until I both try them out myself and, as is the force of my recent experience, see normal non-programmer computer users try them and like them. As for Mozilla, I recently installed it on Windows and it is frustrating to me: 1.) When searching mail messages, the search results windows in Netscape 4.7x had a <Go to Message Folder> button that permitted me to view the message in its folder context. Very useful for viewing a found message in the context of a threaded folder. Since I have thousands of email messages classified into folders (for dozens of projects I have worked on or am working on) I need the ability to jump to the threaded context of a found message and I see no way to do that in the new Mozilla mail client. The mail search results in Mozilla mail no longer offers that ability. The only options are: <Open>, <File>, and <Delete>. Note on folders: Folders are an anachronism. We need a metaphor of 'classifying' objects into a lattice (directed acyclic graph) that does not force us to 'move' things into 'folders'. The lattice semantics would be a superset of folder semantics but, basically, you would keep all of your mail (and, for matter, all documents) in one 'Inbox' and the view restricted subsets of data via filters on the classification lattice. The filters would be expressed in a query language having the power of SQL or first order logic. In this way one mail message (object) would be able to be 'stored' in one place but be represented in multiple places ('folders' in today's jargon) which would have EQ copies of the object. This notion generalizes to permitting relational modelling of objects, getting ride of files altogether. I figured I may as well throw this idea in here for the heck of it. Now THAT would be innovative User Interface (and application) design that moved away from the Windows paradigm.
3.) I can no longer (compared with Netscape 4.7x) copy and paste from the headers part of the email message. They appear in some kind of speical windows that is mouse senstive for adding to address book but not mouse sensitive for simple copying or cutting text! Netscape 4.7x permitted one to cut and paste from the headers. In general, *every* place that shows text on my screen *should* permit me to cut or copy and paste that text (even title bars of windows). Again, *that* would be a nice generalization and improvement to the Windows paradigm (and *very* hard to design consistently and implement -- I do not blame the Free Source community for not thinking of or not bothering to come up with such innovations, yet of course some day someone will come up with such ideas or even better ones and that day will be the day progress really occurs). 4.) Mozilla is slower than Netscape 4.7x (on Windows, haven't checked on Linux yet). I'm sticking to Netscape 4.7x for now (on both Windows and Linux). I *hope* Mozilla will succeed. I will start using it once the email composer permits me to compose in plain text and correctly does word wrapping and insertion of included text and, in general, the system simply *feels* fast and clean.
Maybe Netscape has grown too big for its britches and is doomed to be bloatware. I have too many thousand messages archived in my IMAP folders and not enough time to research a new mail client just now, but I may be forced to if the trend continues. Maybe Galeon or Opera or another browser is the way to go. I've tried Opera a couple of times and it was not ready yet (four years ago and about two years ago -- still didn't have Java applet support). But maybe those are closer to ready now. I am open minded enough to try them at some point (if only I could find the time) but I dismiss the religious zealots who blankly state that they are THE WAY. Bull ####.
> Your printer-example only shows that official drivers I must take this moment to make one of my general remarks. It has *everything* to do with linux. The issue is technology *culture*, not bits. UNIX has has problems with printer drivers since time immemorial. I remember back in my research days how often the postscript printer would screw up and the sysadmin would spend time getting the printer driver to work. I am NOT blaming linux or Torvalds or anyone for the fact that it is HARD to track all printers and keep printer drivers up to date. I am simply pointing out that linux (in the large sense of the term) is not ready to compete with windows in this regard and, ergo, millions or people will not be happy using linux until such time as they can plug in their printer and the printer will work. How can such a statement be controversial. Maybe I'm wrong. Maybe, as one poster mentionned, there is some new printer driver program (CUPS was mentionned) that solves my problem. But my newbies don't care about CUPS. They care about printers. No one who replied actually claimed that they gould seemlessly configure a LaserJet 6L in Madrake or with CUPS, so the case is open for my newbie in question. I'll try again next summer and revisit this. Things *are* improving, of course. Hell, I was just happy that my brother was able to get his Deskjet print working at all, compared to where Linux was four years ago!
The issue is one of *interoperating* with Windows artifacts.
Since so may people claimed that KOffice can no do that so seemlessly, I will certainly give that a good look in my 'next summer' review I am hinting that I'll do. And this time I'll write a thourough real article. (promises promises.... but where to find the time?
> Some things are not yet as far evolved as with windows, For stand alone use, yes. For use in the real world, not yet.
I already made one general remark. That the issue is not micro, but macro -- the issue is about the larger concept of the Linux OS and how it interoperates with people's real needs in the real world which, unfortunatlely, requires Linux to interoperate with Windows artifacts. It has a ways to go before it gets to a satisfactory level. Another thing I noticed was that the vast majority of the responders seemed to be technically savvy programmers or power users. Although the responders make many good and valid points, I take some of them with a grain of sand. Hey people, *I* can and *do* use Linux as a Desktop machine. My 'article' was about some newbies who, (a) needed my help to even begin to use Linux and (b) were frustrated by the experience. Many of the responses stated what is *good* about this distribution or that office suite, etc. My points were not addressing what is good. There are plenty of things I skipped over about what is *fabulous* in Desktop Linux. I was concentrating on problems that *did* and *do* exist for newbies and I am sure that similar problems exist in each of the claimed distributions and suites, etc. Telling me X does Y good does not clue me in as to what X does bad. The issue of a common widget layer is not a trifle. Windows has the luxury of having kernel DDLs built in to a layer that applications can access with no concern for standards. They *are* the standard in the Windows world -- a *bad* standard but, nevertheless, a de facto standard. (see: http://linuxtoday.com/news_story.php3?ltsn=1999-08-04-012-10-NW-SM ) For the Linux community to succeed in establishing cross application consistency for the user experience when using linux, some common widget layer will have to exist. Finally, let me conlude with my pipe dream (well, one of my pipe dreams)... A true move away from the Windows paradigm will involve some future user interface concepts that are not in Linux yet but for which UNIX provides a basis. Do we really need *a* desktop machine? Programs run on servers all over the net. I should not have to run programs on one machine so my user interface should acknowledge that fact of distributed computing.; My bookmarks and address book and other data should be serverized - not on *a* desktop machine. Moreover, why should my Word processor run on *my* desktop machine? The Web has proven that distributed computing is the way of the future. A 'Desktop' should merely be a *standard* interface to distributed computing objects, much as a web browser is a standard interface to HTML via HTTP. Once X-servers have sound built in (I don't think they do yet), once address books, bookmarks, and other documents are serverized and securely accessible over the net (and don't tell me about LDAP unless you've tried to configure your Netscape bookmarks on an LDAP server), then as long as my remote programs interface to (future) X, then I should be able to sit in front of *either* a Windows workstation or a Linux box that I do not even *own* and be able to run programs to use my data in a secure and robust manner. Morale: Down with the Desk Top. Up with standard interfaces to distributed computing. |
RE:Linux not ready for prime time on the Desktop (modified 0 times)
)