Please, take the name 'news' lightly [in addition to the material on this page]. I'm writing this over the course of a few weeks as I learn about these things... Again, think of it as a preview, a sign of things to come.
Well, I finally have a free and useable word processor in Linux. It's called Maxwell and, though, the authors say its in alpha development, the version I have is quite stable and has many features already, including the ability to import and export RTF formats, something I sorely need in a Linux word processor, with partial Word 6 capabilities, also very useful. I've read the copyright and licensing terms. Unfortunately it is not under the GPL yet. The three authors are conceivably debating this. Make sure you check out Maxwell's home page (link at bottom), download the program, and send some feedback to the authors.
Well, now, with this word processor, I can write all future web pages in Linux (I use an RTF2HTML converter for much of the code). Oh, and I can write essays and research papers too. The only thing left is an HTML editor. I've had some success working with HTML Writer under Wine, but I'm missing a few files right now so I can't just do it. I tried AsWedit but it wasn't what I was looking for - I'm very picky what I code HTML with. But guess what I discovered on one of my Linux CD's? Amaya 1.1c! Well I haven't had the opportunity to download Netscape and have needed a stable and standards compliant for a browser for a while, so Amaya helps me kill two birds with one stone as an HTML browser slash editor.
I also found that I had a pdf viewer on one of my CD's. I've been looking for one for a while. This just means even less time in Windows and more in Linux!
In the future I only need to do two things under Windows: 1) print; and 2) HTML conversion [with RTF2HTML] and cleanup [with WebWerx]. I figure this amounts to about 2 hours a week. Gee, it's almost like I'm treating Windows like it's something harmful like drugs...
Guess what I saw the other day? I was at Software Etc. and Electronic Boutique, and they were both carrying Red Hat Linux 5.0! Interestingly enough, it was the only Unix-related product they had...
I've also heard that Best Buy has been carrying it for a while and, now, also CompUSA.
Well, while I was out, in addition to the above pleasant suprise, I bought a Java book and a NetObjects Fusion book. It'll simply be a matter of time before I exhaust both of these resources [and start looking for more?]. Yes, I'm finally going to get into programming! My current plan is to learn Java, then C++ to be able to fix bugs in the software I work with and to write a device driver or two that I could really use. Besides, I've learned as much as I can about Linux and computers without being involved with programming. And no, I didn't get a Java book that includes Microsoft's Java-incompatible Visual J++ (I had the Professional edition before but got rid of it, but that was before learning that it was not compatible with Java). Instead, I have the privilege of using Symantec's Visual Cafe Lite and, of course, the JDK. Learning Fusion should be fun. Maybe that new web site I've been wanting to do for a while will get done after all. (I have FrontPage 98, but FrontPage Explorer insists on having TCP IP installed to even start, and I would take a largely unnecessary performance hit).
So what if it won 3 operating system awards this year?
Linux easily outperforms Windows NT as a file, print, web, and e-mail server. (All at the same time on Linux with half the hardware than NT while running only one of these applications?) But enough on Linux advocacy... NOT!
Hmm, it is, sometimes considerably, less expensive and easier to administer and configure (try doing the same task with Red Hat Linux 5.0 and NT 4.0 and compare for yourself). Linux' ability to communicate with other computers via you-name-it-Linux-got-it protocol, without really expensive software, is arguably one of its strongest points, even when compared to commercial Unix implementations... more on this later. However, I've found that transforming Linux into a desktop OS [as opposed to the client-server model] like, err, Windows 95 requires either: 1) more or less money (I personally tend to go with 'more;' but you can achieve more or less (that expression again) the same result with 'less'); 2) a high-speed 'Net connection; or 3) a super-human amount of patience. In addition, of course, it requires a bit of effort on your part. I think it's actually easier to set up a Linux server than a Linux desktop at this point.
And what else you ask? Linux is probably currently the best platform to run and develop for Java at the moment, but who knows how much that will actually matter?
Caldera has released OpenLinux 1.2. The 'Standard' version now costs only $200 and not $400 - still too much for a commercial Linux distribution if you ask me, but considering Caldera's 'official' competition, this price may even be considered, at least more, practical. Their press release was less than informative as to the specifications of this new version. But a fairly recent update to their web site makes me reconsider Caldera's distribution. Caldera's Netware client is now included in the affordable base package. StarOffice 4.0, BRU PE, the Looking Glass desktop, and a graphical PPP config. tool are also included. This uses kernel 2.0.33. Again, of course, Caldera has not mentioned the distribution COL is based on. My guess is Red Hat 5.0.
Note: If you are looking for an inexpensive Red Hat Linux derivative, then WGS' Linux Pro is the way to go. Unfortunately, glibc2 is [in] the future of x86 Unix, and, at the time of this writing, WGS had not released a Linux Pro that uses glibc2 as the primary C library.
Solaris has extended its upgrade program and decreased the price by 70% for Solaris. This is undoubtedly targetted at Windows NT reading from the press release, and it seems that licenses for Solaris x86 now cost less than licenses for the less stable and technologically inferior Windows NT. Sun has included Linux as a platform from which to 'upgrade,' by the way. It is also probable that there are educational discounts for Solaris [that make Solaris even more affordable for students], but I have not confirmed this. Anyway, this recent announcement is only a good thing for Sun, whose main commercial competitor on the x86 is SCO, as it can only promote the use of Solaris on the x86 [as opposed to SCO's reportedly worst Unix implementation(s) everTM which is currently dominant in the commercial x86 OS market].
Update: it seems SCO OpenServer's file system organization has been really messed up since at least version 5.0 at least partially due to a requirement that workstations running the OS could boot diskless.
Cygnus' GNU-Win32 beta 19 has been released at last. This is a major update. Please inform me how performance is under Windows if you decide to install it. Should I switch from Unix for Windows?
I can't believe it, but Millikin U. is offering me a scholarship. It's only $10,000 a year so the rest will have to covered with loans and work, then my parents' contribution. Still, that's okay. I'll have to see if VU will readmit me before I make a decision, however.
It's nice to see that my work in high school is still paying of, but it upsets me when I encounter an all-too-typical student from my former high school who values his minimum-wage job and his social status over his schoolwork by far. Neither of these will matter to him less than four years into the future, but he doesn't know it yet, and he doesn't bother to do more than show up for class if he does that much. This is the "normal" student at my former high school. The honor students, however, at my former high school are easily distinguishable by their arduous studying five minutes before any given test or copying from a classmate the previous class hour when a homework assignment is due. Then again, the teachers and the administration (and even parents), in general, don't fare much better. Their most obvious fault to me is the treatment of all students as deviants, not unlike the treatment of prison guards of their prisoners. They are as defeated as the students are in their thinking.
I personally have been distracted as of late due to a few, but significant, personal problems. Unfortunately I have two tests this (err, this has long past) Thursday and an in-class essay in two weeks (still have to do this). The latter involves visiting the local historical museum beforehand. I'm also [finally] getting a job, hopefully one that makes use of my age, high school diploma, and, if possible, my computer skills. I'm just wondering what kind of car I'm going to get (I've been driving my mom's car - a '94 dark evergreen Grand Am SE)...
The most important news I've heard is the release of SCO Merge 4.0. This update of Merge will allow one to run Windows 95, Windows 3.x, and DOS applications in a Unix environment. It is free for individual and non-commercial use and can be downloaded from SCO's web site.
There are several downsides to this. The first is that it runs only on SCO's OpenServer 5.0.4 or higher. Initial attempts with iBCS under Linux have perceivably failed because of missing syscalls in the Linux kernel. Second, it is very much more like DOSEmu or Bochs than it is like Wine or WABI. This means that Merge emulates the hardware and does not attempt to implement the Win32 API like Wine does. This hardware emulation will result in a serious performance penalty, not unlike VirtualPC under Mac or SoftWindows95 under Unix. And there are other limitations as well. These are due to the 'limitations' of Unix systems so you will not be able to run many Windows 95 games as well as a predictably few applications. And last of all, since this is emulating a PC and not Windows, a licensed version of Windows 95 will be required.
Err, thus said and recognizing the limitations of Merge, it looks like I'll be installing Windows 95 again after all, but at least this time it will be under Unix. It would be great, however, if the Linux developer community somehow got Merge to work under Linux and iBCS2. I am saying 'community' because I believe this may take some serious hacking. Of course it would be better if Platinum Software(?) ported it to Linux. (Could someone give me the number of Linux installations compared to SCO Unix installations? It may make a convincing argument.) But what would really be better is if Sun, or err, even Caldera, continued to work on WABI or at least release it under the GPL. But what's the likelihood of that? Probably the same as that of Microsoft releasing Windows 95 under the GPL.
The only question that remains unanswered [for me] is how stable Windows 95 will be under Merge. I hope it's more stable than Windows 3.1 under DOSEmu. WABI is looking more and more attractive, especially since Caldera has just reduced its price again to about $50 (the old, old price was $200).
Update: Actually, I've decided not to get OpenServer and Merge, if only for the above concerns and perhaps upon learning that OpenServer is considered by many to be A Really Bad Unix Implementation (TM). But I am getting WABI once I can find the money.
I've been thinking about getting Solaris (a SVR4 Unix as of version 2.x), but...
Linux already scales better than NT with two processors, the most I'll ever have, and is more stable (I've had a number of lockups in NT just this week that were not my fault as Microsoft would have you believe, but these have only forced me to log out and re-login - almost like a reboot but just a little bit better). SMP in Linux is still considered to be in the experimental stage, by the way. Plus, Linux is more likely to be [commercially] supported on the x86 than Solaris as far as the availability of software is concerned. So, there's really no point to getting Solaris x86, especially when WABI is available for Linux (the only reason I'd get it since I already have CDE - the only other cool thing about Solaris x86).
POSIX support in NT is admittedly broken according to Microsoft. This is, of course, after Microsoft went to court and spent millions of dollars in order to ensure their right to claim that NT is POSIX compliant. I believe they won the case. Well, on the other hand, if it's not true that Microsoft has admitted that NT is not POSIX compliant (since I personally have not confirmed this), then why hasn't Microsoft submitted NT for POSIX certification? They certainly have the money, and it couldn't but help the marketability of NT. So why hasn't NT been POSIX certified? The reason for this, of course, is that NT wouldn't pass the certification tests.
Linux 1.2 and higher is reportedly 100% POSIX 1003.1 compliant. There is at least one distribution of Linux that has been POSIX certified. Oh, yeah, and Linux is free.
NT has a C2 security rating, but that's with no network connection. Yeah, that's really useful (just as useful as a computer with no power cord). The ironic, or sad, thing is that people actually bought it, literally and figuratively.
NT makes the basic administration tasks easy enough that extensive knowledge is not needed. But this also means that in case of problem your average NT admin[istrator] will choose to reinstall the server instead of trying to figure out the problem and fix it. Unfortunately with Microsoft products, it's often the fastest solution to configuration problems...
Since Linux, and to an extent Unix, doesn't try to hide the complexity the way NT does, configuration problems can always be solved. And with the superior documentation and technical support, not to mention source code, problems can almost always be diagnosed and fixed faster.
Do you want to know what Windows NT will look like in the twenty-first century? Take a look at Linux NOW.
I have not done much [formal] research, but these are my initial reactions to the Microsoft monopoly. Yes, think of it as another preview or an introduction to a much lengthier editorial [with references!] [as should everything else I write on this 'news' page].
Well, no one can deny that Microsoft has been introducing more and more features with each release of their flagship products. But it has conceivably been introducing more and more bugs as well, and bloat. But, at the same time, they have been stifling technological advances purposely with the assistance of, and in order to preserve, its monopoly. Just think about OS2 vs. Windows 95, or now, Windows NT versus everything else. Think about OpenGL technology, ActiveX, and all the other technologies that Microsoft has promoted despite these technologies inferiority to [now formerly] competing technologies. And they have been overtly successful.
Netscape has probably presented the [only] answer to the Microsoft monopoly [at least to the rest of the world that doesn't know about Linux and open source software]: free software. Microsoft cannot stop free software. The people who do free software aren't being paid for it, therefore Microsoft cannot control them. The GPL has, in part, allowed these people to develop software [over the Internet] without the fear of having their work made propriety. These 'people' tend to be some of the best programmers in the world, and because they are not under a single company's control, they are free to do what they want. And because of this, they are willing to push technology to the edge; they do not benefit from inferior technologies as Microsoft many times has.
No, I don't like the BSD license. Any programs under this license can be made propriety [and source code no longer made available], meaning development can cease if or when interest wanes, or if the commercial entity in charge of development chooses to cease or to prevent further advances that would otherwise be beneficial for the community as a whole.
I know I'm a little late on this 'new' technology but better late than never. I'm a little disgusted with the term 'new,' because dumb terminals have been around for a very long time [so I'm told]. These are basically what NC's are, just without graphical interfaces.
This said, why not just turn all those Windows, OS2, or MacOS boxes into X terminals? Same thing, but you don't have to buy new equipment, and they will just be as easy to administer as NC's. Even boxes with as little as 4 MB's of RAM can be used as X terminals as long as Linux is running on them. Need Windows applications? As long as you have a server capable of running the binaries then you can do that too. Hmm, I don't think this is a 'new' idea either. A number of companies are evidently making a living developing and selling X servers, not to mention terminal emulators.
What about having web browsers as a common interface [to reduce administrative and training costs]? Wait, that's not new either. Isn't that the basis of intranets?
Now this gets more and more interesting as I hear more about it. From what I've read this is going to be one heck of a development environment. Or is it that it has been a heck of a development environment? At least Rhapsody won't cost $20K to evaluate like NeXTStep slash OpenStep did. With the Yellow Box technology it appears that everybody, meaning users, not corporate entities, will benefit from the success of Rhapsody. There's something to be said about platform independence. I know, there's Java, but Yellow Box, from what I've read, is actually practical (this would be very much a rehash of the debate of Linux vs. Hurd, which is still not finished).
Still I have to frown on the recommended 48 MB of RAM to run Rhapsody. More significantly there's no initial SMP, built-in X Window System, or official POSIX support. SMP is a necessity if Rhapsody will ever be used as a server, a use for an OS that has been proven to be very lucrative. And though there's no X, Rhapsody does allow one to choose between a MacOS or the NeXT based interface. But this is hardly a substitute for the configurability or flexibility of X, and the dual-interface has the downside of adding to Rhapsody's relative bloat. However Rhapsody will almost undoubtedly have better POSIX support than NT because of its BSD roots, but the lack of official POSIX support will only discourage its use in government and government-related agencies, a fairly significant market.
I think Power Users, Apple's conceivable target with Rhapsody, are more likely to benefit from BeOS. Don't ask me why yet however as I'm still doing research to support this contention. OpenStep slash Rhapsody appears more to me as a developer's OS (are these people the Power Users?). However, I'm not implying that Rhapsody and BeOS can't coexist [or that BeOS isn't a good development platform]. I'm sure there are those who are attracted to Rhapsody's Yellow Box technology, and there are those attracted to BeOS' multimedia capabilities.
Yellow Box is the biggest attraction with Rhapsody. Unfortunately, it is obvious that Rhapsody must succeed first before Yellow Box can and not the other way around. No one will develop for a platform that doesn't have enough 'seats.' This is one thing that Be evidently knows that Apple has perhaps knowingly or unknowingly ignored: before people will develop for a platform, the platform must have a relatively large number of users. What is an application without users after all? This is basic demand and supply economics. I wonder what the average level of education is for a CEO at Apple... Apple better start to do more in promoting Rhapsody if this OS is going to go somewhere except down in history as Apple's worst, and maybe final, mistake.
To somewhat conclude Apple seems to be taking a really big risk with Rhapsody and Yellow Box. Only time will tell if their gamble was worth it. (Have the people at Apple made any good decisions? I don't recall having heard of anything positive that they have done to promote their hardware and software).
Update: Apple may or may not have announced the availability of Rhapsody and Rhapsody development tools to universities for no charge. Additionally I have learned that though perhaps OpenStep, therefore Rhapsody, is derived from BSD 4.4 Lite, like FreeBSD 2.x or NetBSD 1.0 and their derivatives, OpenStep is more appropriately related to Mach 3.0 (MkLinux is based on Mach 2.5, btw, and some of the programmers who worked on MkLinux are also reportedly working with Rhapsody, whatever that means). So it's somewhat unfair to call Rhapsody just another BSD Unix.
I actually had a few problems with installing CDE, but they weren't really problems. The setup script was a bit less forgiving than I could have hoped, improperly writing a line in ld.so.conf. But I think that was the only thing I can complain about. All the other 'problems' were rather related to the limitations imposed on X by CDE for evidently security purposes and the expectation of CDE to be connected to a network. The latter has only [considerably] delayed startup after login and has required me to run daemons that I would otherwise not have to run. I'm not sure which daemons are absolutely required (I have not been able to find the appropriate documentation) for CDE to start, but both of these problems are tolerable annoyances. However, I'm quite impressed with performance on my P166 with 32 MB's of RAM. Performance is comparable to the Sparcs I worked with in the engineering computer lab at Valparaiso University.
I like the CDE. First, it makes my Linux box look really expensive. But more than that, it's quite user-friendly, more so than any other [propriety] desktop I have worked with. There are applications included with it, of course, namely a text editor, a terminal emulator, a calendar, a calculator, an e-mail client, a file manager, etc.
However, I think the KDE is even more user-friendly and perhaps has more potential, because it's somewhat free as opposed to the commercially restricted CDE. Maybe when it, or GNOME, is no longer in beta I'll switch to it. But in the meanwhile, I'm using CDE because it works for me now. Besides, I've spent a pretty penny on it.
InfoMagic's LDR is really something. If you want to get into Linux, this is definitely the kit to buy. The January release includes the three most popular distributions: Slackware 3.4, Suse 5.1, and Red Hat Linux 5.0. It includes files from the tsx-11, sunsite, and GNU archives, practically all the freely distributable Linux software available. It has XFree86, but if XFree86 doesn't support your video card, then maybe the full commercial version of Metro-X included in LDR does. This is really one of the best reasons to get the LDR. Separately, Metro-X alone would cost almost four times that of LDR. The most recent kernel sources, Apache, Lesstif, the Red Hat Power Tools, and, of course, all the documentation from the Linux Documentation Project are also included.
The only thing missing from the LDR is printed documentation. But this need is better met by Running Linux and Linux in a Nutshell.
Suse has enhanced XFree86 [for use with Matrox Millenium?] Go Suse's web site for details.
Well, it appears Debian 2.x will be impervious to the year 2000 problem. But here's a quote from the Debian GNU Linux - Latest News page anyway.
Unix and Linux do not store 2-digit dates, although it is possible for an application to do so. They store dates as a count of seconds since New Year's Day 1970. This counter will overflow about 40 years from now, in early 2038, _not_ 2000...
Before 2038 we must define 'time_t,' to be a 64-bit variable instead of a 32-bit one, and recompile all programs. This is a very simple process compared to the anguish the non-Unix world is going through - we go through more work to produce a major release of Debian. Once time_t is a 64 bit variable, it's good for roughly another 292271023017 years. By then, there may be something better than Debian :-)'
Well, I don't think I'm going to hold out that long until I get the next Debian release. But I'm really looking forward to future releases of Debian. Glibc2 slashlibc6 is definitely the latest thing right now in Unix development. But what's nice about Debian 2.0's use of glibc2 is that Debian has openly stated that they will maintain compatibility with libc5 applications. This is A Good ThingTM.
Debian 2.0 has been in development since last July, 1997. As of February 3, 1998, Debian 2.0 is 'near completion.' The '2.0,' of course, refers to the use of glibc2 as the primary C library, a significant change that warrants a major change in version number.
A replacement-supplement for dselect will be released shortly after the release of 2.0. This can have a X based interface, but it can also be accessed via terminal. It almost sounds like glint and rpm, but since Debian packaging technology is superior to rpm, maybe the new dselect and Debian packages may catch on, and Red Hat will do the right thing and adopt the Debian package format. I personally believe the Debian version should be upped to 3.0 once this new dselect is released, since I had such a bad experience with dselect that, instead of learning how to use it, I opted for a different distribution.
Version 1.31R6 has been made available by Cheapbytes on CD for the price of $2.49 plus shipping and handling. This will probably the last non-glibc based release of Debian/GNU Linux. 0.93R6 was the last a.out release. The reason I mention this version is that:
Until the release of the new dselect, I will probably not even try to install Debian. Other things that would be nice to have are:
I know, I'll be waiting a long time. But Red Hat is currently quite satisfactory, and I see no reason to switch (I mean, buy the CD's and reinstall Linux) until these programs are released. The lack of a year 2000 problem in Debian 2.x is probably a good reason to switch, however, before that time (while the present kernel will hold up to 2038, the Debian development team has been working on the applications included with Debian which presently, with 1.31, do not).
If you or someone you know requires a computer lab assistant in the St. Lucie County area, don't hesitate to contact me. I have experience with both DOS and Unix-based operating systems. I have formal training in business office applications (I type an avg. of 70+ words/minute), and I possess general office skills. I have an extensive knowledge of Windows software in general, but I prefer Unix and editing text files to frequent crashes ;-) and inaccurate point-and-click software configuration. I know HTML, of course, up to and including 4.0 extensions (yes, a complete rewrite and redesign of this site is in the works - the current 'site' was designed one month into learning what the acronym HTML stood for, so please excuse the poor design) and have a familiarity with almost all web-related software, including a variety of web servers. I know how to locate information on the Web in addition to promoting a web site. Computer technician or even systems administrator isn't beyond my league, but I will require some training [either formal or on my own]. A volunteer position in a non-profit organization is acceptable if not preferred, but the position must be related to both my interests and skills. A formal resume will be given if requested, and an interview will be granted if I find the position desirable. Oops, by the time someone reads this I'll probably have a job (hopefully one that requires some level of education).
Need an operating system? If you have less than a 386 with 2 MB's of RAM, go with Caldera's DR-DOS; otherwise, go with Linux. Linux' weaknesses are fast disappearing, and the quality of this OS and the license it is under ensures that it will always be supported. DR-DOS is free for non-commercial use; a commercial use license is only $30, but this DOS is a 100% compatible DOS. Multi-tasking and modern memory management, as well as built-in Netware support, make this the best DOS. Caldera promises Java support for DR-DOS by the end of the year.
DR-DOS is basically Novell DOS 7 with patches, a year 2000 work-around, and other enhancements. It was formally known as OpenDOS, and for a short while, as DR-OpenDOS. The kernel source code is available, but this is not under the GNU GPL. Support for a dual-boot with Windows 95 was implemented in version 7.02, and long file name support is currently in development(? possibly finished already?).
The first two are links to CDE front panel clones. These are really close, and with a little effort one can 'create' a CDE look-and-work-alike. Both are optimized for fvwm2, I believe.
Copyright © 1997 Mark Santos. All rights reserved.