The collected works of wenp

jeremy123 wrote: I have tried aix 5 - 7 very interesting os...

It sounds like you tried AIX 5 to 7 on the same machine. How did they compare in terms of performance and resource requirements?

I knew that IBM licensed hardware per core on mainframes, but this about POWER is new to me. Does this mean, then, that if you buy a used POWER6 you then also have to pay IBM to get the cores activated? Is IBM the only maker that operates this way?

A couple years ago, I read that Oracle's predatory pricing and stalled development of SPARC was driving clients to IBM. I wonder if IBM is now driving some back.

OpenPOWER looks interesting, but if you don't have AIX, what does it have over a new Xeon? I'm genuinely curious about the details.

Kira wrote: [*]Power8 is really fast

Just on speed? I was sort of expecting that hardware support for virtualization would be a factor. Is PowerVM a big advantage, or can you approach similar performance with Xeon?

On old 20th-century machines, I've sometimes had to revert to Squeeze and kernel 2.6 for good performance, but I doubt you want to do that. For convenience and stability, I think Debian is still probably your best choice. The Red Hat derived distros have a heavier basic configuration and are hard to slim down. The RPM-based package ecosystem is very limited and awkard compared to Debian.

I've tried just about every lightweight window manager out there and found one of the most stable to be IceWM. When users want a Windows-like interface, IceWM is what I give them. It is IMHO the most user-friendly of the light WMs and there is a great range of themes available. For the Windows refugees, I make it look like Vista. It does take a lot more manual configuration than setting up KDE or GNOME.

My taste is for tiling WMs, but the ones I tried were relatively unstable and not particularly lightweight. Since I found I could configure CTWM to look and act pretty much like a tiling WM, it has become my default choice.

hamei wrote: We should face facts : in 2000 A.D. Linux on the Desktop was a nifty idea.

I completely agree with your evaluation of Linux, but I think the OP's case is probably best served by Linux. When I set up an Internet machine for the old folks, I use Debian. It supports just about any random cheap/old hardware and modern browsers, and doesn't require constant vigilance to remain functional in the face of attacks. I fantasize about using OpenBSD in that situation, but the hardware/browser support will never be there.

There were a few times when I said to myself, "Windows can't really be so bad if you know your stuff. I'll just get myself some really thick books on the subject, study hard, and learn how to achieve security and peace of mind on Windows." On the rebound, even Linux looked elegant.

So, it's well known that a beta was released and that the project was killed after selling less than 50 licenses.

The Wikipedia article seems to give a good explanation of how IBM, SCO, Intel, and a few others were trying to build a unified Unix that worked across POWER, IA-64, and IA-32, but were undercut by late hardware and the rise of Linux.

What I'm wondering is how much of the work done made it into other projects? It looks like all the Itanium related stuff was abandoned.

Good collection of IBM docs at:
http://ps-2.kev009.com/

Device driver kit (and a bunch of other old stuff) at:
ftp://ftp.software.ibm.com/aix/itanium/developer/

Evi Nemeth, original author of the famous Unix System Administration Handbook , was lost at sea last summer. Something that was commented on by those who knew her was how much she liked HP-UX and hated Solaris. While the Solaris bit wasn't unusual, I'd never heard of anyone loving HP-UX. I just spent some time comparing editions of her book to see what her issues were.

Reading USAH 2nd (1995) edition, yeah, she really complains a lot about Solaris. Her big things are disorganized init/log facilities, tools with odd names/features, complicated networking config, and flaky hardware. Nemeth says the modular kernal architecture is nice when automatic driver loading works, but is tortuous when it doesn't due to frequent hardware problems.

On the flip side, Evi thinks HP-UX is just brilliant, with hardly a negative thing to say about it. The man pages are much improved from careful editing. Init facilities are neatly organized. Hardware generally "just works" (due to the limited range supported?). Networking is super easy to configure and rock solid. If there was any feature she didn't give a complete thumbs up, it would be that booting procedures tended to vary by machine.

Going to the 3rd (2000) and 4th (2010) editions, she vented less spleen against Solaris, but I can't tell if Solaris improved all that much or she just got mellower. In the 4th edition, I think she was more focused on disliking AIX and its ODM. Solaris wasn't the only focus of her ire. RHEL gets criticized a lot, and she sometimes makes FreeBSD sound like an affliction.

SAQ wrote: In the red version she was noticeably pro-BSD and anti-SysV.

In her evaluation, she seemed to put heavy emphasis on how nicely the OS supported the hardware. In the time of the 1st and 2nd editions, would BSDs have been mainly on VAX? With a limited set of hardware to support, I suppose there would be relatively few problems.

In the 3rd edition, she takes some pains to excuse the extra work required by FreeBSD, saying it's hard to support PC hardware.

ClassicHasClass wrote: ...there is no longer any performance gulf of substance between xlC and gcc anyway.

When I read this bit in Mendoza's UNIX to Linux Porting: A Comprehensive Reference , I was surprised that XLC made so much use of the GNU resources.

...the XL C/C++ compilers employ the same GNU libraries, linker, assembler, and binutils as GCC...optimization routines have been developed for and tailored to the POWER architecture. High-performance computing applications, especially those heavily reliant on floating-point operations, often benefit significantly from a simple recompile with XL C/C++.

Alver wrote: I honestly love UX, and still use it at home for a desktop

As I noted, a big part of Nemeth's affection was the reliability of the hardware and how well the OS supported it. By a wide margin, I hear more praise of the reliablity HP hardware than any other Unix boxes, but that's mostly from the PA-RISC days. Are things still as good with Itanium boxes?

Just some historical comments from Bennett's Inside the IBM RISC System 6000 , available at http://tinyurl.com/k7lb5rv

This is IBM's version of UNIX for the PS/2. It is the result of a joint development between the California-based Locus Computing Corp. and IBM. AIX PS/2 was originally based on a highly modified version of UNIX 4.1 BSD.

AIX PS/2 is positioned as an entry level to the IBM AIX family. How well does it live up to this claim? IBM has had only moderate success in marketing AIX PS/2 as the UNIX system of choice for personal computer users. Why? First, AIX PS/2 predominantly runs on IBM PS/2 computers and provides only limited support for IBM compatibles. This is because AIX PS/2 cannot use the IBM Personal Computer BIOS which is written to the requirements of a single-tasking operating system like DOS. Instead, AIX PS/2 addresses the hardware directly, and because other vendors' non-MicroChannel bus-based machines have various hardware differences they will not necessarily work with AIX PS/2. Second, in the personal computer UNIX marketplace there is no de facto agreed standard for binary application compatibility. The marketplace for AIX PS/2 systems therefore is typically a small configuration, multiuser commercial system. For example, a customer may develop a simple character-based application on an IBM RISC System/6000 for its worldwide series of offices. Smaller offices may not justify their own IBM RISC System/6000 so the application could be ported to an IBM PS/2 running AIX PS/2 for these smaller branches.

Ironically, since the cost of the IBM RISC System/6000 is now very low, it can outperform a high-end IBM PS/2 in terms of performance per pound spent, thus making the choice of AIX PS/2 more difficult.

vishnu wrote: I have diddled around with the fox toolkit, but gave up on it because it couldn't (at least not then) handle UTF8.

The FOX toolkit does support UTF-8 now, just not RTL text.

I favor simple code over visual bling, so if I liked C++, I'd be using FLTK. XForms seems to be the simplest toolkit in C. It now has beta support for UTF-8.
http://xforms-toolkit.org/

You might want to look at IUP. It is not just an elementary toolkit; there are many powerful widgets. The user community is so small I suspect they could all fit into a bus, but the developers are active and very responsive. On Unix/Linux, you have a choice of GTK or Motif backends, and they actively support some Motif based apps on commercial Unix. You can write with Lua or their own LED, but I found it unnecessary since the C code was so clean.
http://sourceforge.net/projects/iup/

First Spock...er, Leonard Nimoy...now Pratchett. Really enjoyed his earlier ones. Probably the best introduction to the series is the first book, The Color of Magic. For me, the funniest was Pyramids. Some of the entries with stronger plots, Sourcery and Reaper Man.

guardian452 wrote: hmmm, slight addition:

8% run more than one operating system on the same machine
50% have a Windows PC desktop or laptop
58% have a Mac desktop or laptop
5% have a Unix or Linux computer

Curious how many of those 8% are virtual machine vs. dual-boot or boot camp ?

That survey tells us a lot less than it appears to, since it doesn't give a breakdown by major. I'd guess that nearly all *nix users are Engineering or CS majors, and that most humanities majors are on Macs. So it probably hasn't changed from the height of the Linux fanboy epidemic.

I started with Linux to get away from Windows 98. Many years later, I hate going back to Windows because I lose the advantage of hard-gained experience in Unixy tools and methods. Yet I spend half my time on Windows machines. I used to (mildly) proselytize for Linux, but I just realized it has been years since I recommended it to anyone. If I were a student today, I would investigate Linux (and all the free *nixes), but I'm pretty sure I'd make an informed decision to stick with Windows.

commodorejohn wrote: The only problem with that is that Microsoft keeps kneecapping Windows by gradually eroding its advantages

Heh, heh. Still using WinXP when I can.

...prospect of being able to just set myself up with a single software environment and then never have to put up with stupid UI changes again.

I've finally accepted that's never going to happen. Even if I make a comfortable setup in NetBSD or OS/2, it's like living on a desert island when practically everyone I might want to exchange software with is on Windows. I've basically resigned myself to using software whose design is repulsive but objectively allows me to get more work done. Is that wisdom or the defeatism of old age?

Still, when I think about coding my projects just for Windows, I become really, really sad.

TeamBlackFox wrote: ...on OS X it is a clusterfuck of garbage design, launchd, Mach, Aqua, and it can't even decide if its GNU, BSD or Darwin

[with a glance in the rearview mirror at the OP topic falling behind]

People keep telling me OS X is Unix with a polished desktop ecosystem. Sounds wonderful. And when I first saw someone typesetting mathematics on a Retina display, yeah, I heard angels singing. Sometimes I think I would buy a Retina MacBook just for that one use case.

Unless you are counting Android, though, I'd say OS X is the least Unixy Unix-like system out there. You have a case-insensitive filesystem, no /proc, different organization of system libraries, a unique method of linking shared libraries, an entirely different I/O framework, and scant control over the interface. X11 apps are very unstable on OS X -- no clipboard integration, mouse driver locks up, spontaneous loss of focus.

Sure the development environment is very integrated, but also rigid, slow, and bloated. Forget about cross-platform development unless working entirely in a VM. The range of dev tools is limited compared to Unix/Linux. In general, OS X has much fewer applications ported to it than Linux or FreeBSD.

The OS X API has been no more stable than Linux and not as well documented or as neatly organized. (Hard to believe I just used Linux as an example of better organization.) The community is just as dominated by fanboys as Linux was at its worst, and in my brief experience no more helpful on technical matters.

[whoa...glanced again and I see the OP tailgating me now]

I should qualify my rant by saying that I develop apps for workstation use, and practically everyone I care about is on Windows. So for me, using another system as a dev environment is purely for my own satisfaction. Eventually, I have to build everything for Windows.

foetz wrote: windows is a joke and serious software support for linux never happened so hello osx

The relative stability of Mac and Windows has long been a contentious issue. In the days of Mac OS, I knew a university sysadmin who would brandish his logbooks to show that Macs took twice as much maintenance time. Nowadays, you can find some discussion of the "it just works" reputation of OS X, the gist of which is that when you get down to specific cases, Macs are really not more reliable.

In certain areas of software, though, I agree Macs really have an edge. Graphics and video/audio seem to be much better. As for Windows versus Linux, I have been surprised at some of the things I find myself still going to Windows for. For example, there still isn't a hex editor for Linux to compete with the likes of 010 Edit.

tomvos wrote: His lack of ability to acknowledge other peoples work is one thing I really can't stand.

I find his insistence on "GNU/Linux" funny. Have you ever noticed how many tools in a typical GNU installation are taken from BSD? I doubt he would like to call his toolset "BSD/GNU".

GCC is another thing. It actually originated as someone else's Pascal compiler. It had good Modula-2 support until the GNU developers (of which Stallman was just one) deprecated it.

A table sold as a piece of office furniture will be more expensive than a plain old table. To maximize table for the money, how about a plain utility table or dining table and mount a keyboard drawer under it?

hamei wrote: ...after five years there really is a new version of NEdit

I noticed that the last entry in the changelog is from 2007.

Are you looking only for something that can be built on IRIX? If you don't like GTK, are you hoping for a Motif interface?

To save anyone else the same hassle, I'll confirm hamei's opinion. I tried all the font utilities I could find for Linux, but gave up and now I do all my font editing in Windows. Even the free stuff is more stable than anything on Linux.

hamei wrote: But pdf's look excellent. And they look excellent on the Winders box. Viewing the fonts on Winders, they all look the same - crappy.

Antialiasing on Windows was designed to make text crisp and easier to read, so the shape of characters is squeezed to give a greater proportion of solid black/white pixels. On the Mac, they went the opposite direction, keeping the proper geometric form of curves and allowing more gray pixels. So fonts on the Mac have a nicer shape but tend to look fuzzy.

Adobe Reader uses its own display engine, which treats text more like a Mac. If I have some cheap fonts that aren't displaying well on Windows, one workaround is to put them in a PDF and do a screen capture of the display.

I don't work in printing, so I have no direct experience with this, but I hear frequent mention of the color issue from graphic artists commenting on GIMP, Inkscape, Scribus, etc. Are there any open source tools that handle color to industry standards?

Roger Penrose's theory of the mind gets a lot of attention among computer types, but specialists in the topics he addresses have dismissed his work.

He completely misunderstands the range of application of Goedel's incompleteness theorems[1]. He posits quantum effects on processes that operate at a scale where quantum properties have no significant effect[2].

I seem to recall there were also criticisms of his understanding of biology, but I don't have any references for that.

[1] Torkel Franzen. 2005. Goedel's Theorem: An Incomplete Guide to its Use and Abuse.
[2] Alwyn Scott. 1995. Stairway to the Mind: The Controversial New Science of Consciousness.