hipu wrote:let me drop in my 2 cents, with no particular order to this thread:
thanks hipu, for your thoughtful response. Relica was very helpful with his bash config files and advice on Cygwin as well.
hipu wrote:emacs/vi - yes, you could refer to them as command-line editors, but they are not editors. they are much more than that, with syntax highlighting, highly customization, tags, complex searches and more. ...snip
I'm sure that people who use them become proficient and find them useful. I actually got to a point where I could get around in either, but I did it out of necessity. As I said, they remind me of the old way of text editing, which was aimed more at teletype machines.
Maybe if I was at a university and had to do complex searches and formatting for documents, I could make use of them. I'm just a gronk raised on Dos/Windows, looking for an alternative because I don't appreciate the Big Brotherness of Microsoft. Windows/Dos is just not as bad as it's made out to be, however. It's the unconfigurable addons that are the problem, like getting rid of wbem and other Gates goodies he thought we couldn't live without.
hipu wrote:and about your frustration with linux - oh well..snip...ately i started using virtualbox for the same reason which basically saves me from turning on the xp machine.
I did have the KDE version of RedHat on a virtual box till recently. I'm looking into an external hard drive, so maybe I'll load it again. It doesn't run as fast as it does on a straight drive, of course. The issue is still that I just play with Linux...I haven't found a good reason to dump Windoze yet. Maybe with the advent of Vista, I will likely make the move if XP is discontinued. I'm not going to load Vista.
hipu wrote:The reason i made the switch to Linux (8-9 years ago) was my frustration of Windows. I didn't like the idea of not being in control ("control freaks" tm) which is why i turned into Gentoo. I admired the fact that -i- choose EVERY element installed on my system, and compile it for my machine
that is a positive aspect of Linux, but recent experiences suggest compiling/recompiling can be hazardous. I was trying to fix what I see as a bug in Paimei. It has been suggested the problem lies in my setup but I have done extensive checks and I don't think so. The fact that the author of Paimei can't/won't answer my questions suggests he is having problems understanding his own creation.
Anyway, I started tracing Paimei via Python24 and ran into code that made mfc look friendly. I decided to build a debug version of python, although I had no idea what I was getting into. I don't care what anyone says, the documentation for open source applications leaves an awful lot to be desired.
Compiling Python was relatively straightforward using the free version of Visual Studio 9. It was the addons, like zlib, ssl, wxwidgets, etc. that caused havoc. I spent as much time with wxwidgets as I did with Python itself. It became apparent to me that people writing Linux-based apps for Windows had only an inkling of how it worked. Don't get me wrong, I really appreciate the effort that went into porting Python et al to Windows. All the same, getting support for it can be really frustrating.
During compilation, for example, the makefiles were constrained to using Visual Studio 6, or equivalent. No one could tell me why. Microsoft has been good at keeping their apps backward compatible. The makefiles wouldn't let me use VS 9, so I had to comment out entire sections checking the version. Finally, I got it to compile wxwidgets and wxpython.
I am now in the process of getting Paimei up with the debug version of Python. That's another exercise in futility. It requires constant searching with Google trying to find mods to source code that stalls during compilation for no apparent reason. I came across one fix explaining how to modify source code on an activex module, but it might as well have been in Greek. I deciphered part of it, enough to fix it, but never did find what the guy was talking about in part 2 of the fix. He made reference to code that insiders would have understood but not someone like myself.
That's an issue I have with Linux. As an open source OS, there are many variations of it and many contributors. Subsequently, people tend to write documents as if it is understood what they are writing about. Not so. Trying to make sense of a lot of it, tends to get beyond plain frustration to outright insanity. Google searches are of limited value and forums can be downright maddening with the glib and insulting replies.
hipu wrote:to get the hang of linux you just start using it. you google as you go along, your read the man pages and get the in-and-outs of each command by gaining the experience of other linux users who share their knowledge on the web.
I have no problem with your point of view, obviously no one in RCE learns anything without spending countless hours doing what you suggest. I just haven't gotten desperate enough to force myself into Linux. Part of the reason for that, as I have said, is that Windows is adequate for what I need. Having said that, I'm open to comments and suggestion from people like you. I certainly have no beef against Linux and I am not closed-minded about learning it.
I'm not the type to play with commands to see what they do. With the example I gave above, about compiling a debug version of Python, I had a need to do it. That's what attracts me...necessity. I had forgotten a lot about C compilers, but getting into the command line C compiler again was of interest because I want to learn C at a deeper level.
hipu wrote:in my view Linux is way more advanced than Windows/Max-OSX, in usability, performance, modularity, security, GUI (YES! even gui! see latest compiz?
that's an arguement I'll leave alone for fear of unintentionaly insulting anyone. It's also an arguement that is relative to the level the user uses the OS. There are a few things you left off the list, such as the ability of Linux to crash mysteriously. Then there's the 'hold-your-breath' period as you reboot from a crash to see if the OS will reboot.
My feeling is that Linux is far too delicate, and sensitive to user errors. I guess that's the price you pay for precision and security. I don't need either. I use XP home edition, which thankfully doesn't have the paranoid security features of XP pro.
Win 98 was like that but XP, although not perfect, is quite stable. At least, it is if you leave off the updates beyond SP2, which feature Microsoft's pathetic attempts to fix holes in an OS they delighted in announcing as an internet-friendly OS. They forgot there are people on the net who like to pry into other peoples' affairs.
XP is also configurable to a degree. There are countless tips on the net on how to modify it's behavior, and if you read through the RCE archives, you'll find lots about playing with it at a deeper level.
I remember tracing through ring 0 code in win 98 with softice, expecting a crash at any time. With XP, I find it virtually impossible to crash it using softice. I have no idea what to expect with Linux when it comes to debugging. Most of the apps I play with were written for Windoze.
hipu wrote:Vista and OSX can only dream of such effects, with no bloat and no heavy duty cpu! I was running compiz with full effects (except for rain-drops, for some reason my gfx card didn't like the rain...) on my 1.4ghz cpu with 512mb - no need to mention Vista wouldn't even run on this machine, neither to say with gui effects)
I've been running XP on a 2 gig Celeron for years and it's still fast enough for me. My vid card runs directx 9.0c, which works well for the games I use. I'm still using an old Audigy 1 sound card, using drivers from Audigy 4. I have no idea how any of that would work in Linux.
hipu wrote:and now about the /usr/bin issue - there is a logic in linux/*nix directories. /sbin is for (s)ystem essential programs (ie, programs needed to boot the system. period). /bin is the rest of the system programs (ie, not essential, but needed, like 'ls' for example), /usr/bin is, well... try guessing. guessed? yes, user binaries
I don't have a problem with the logic of unix-based directories (sorry...files
). My problem is the way Cygwin hides the files in the /usr/bin directory and has them linked to the /root/bin directory. While I was compiling, there was a call to the gcc compiler in /usr/bin, and bash couldn't see the file, although I confirmed it was there. I'm looking for a method to break the link cygwin has imposed and remount the /usr/bin directory independently.
i don't believe any of what you've said is a valid point against using Linux. laziness is.
I'll take that with humour.
If you understood what I said about compiling a debug version of Python, and using softice to trace through Paimei/Python code, you'll understand that I'm more insane than lazy.