Samba's Jeremy Allison On Linux's Future
TRNick writes "Jeremy Allison talks Ubuntu, why he loves Gnome, and the trials and tribulations of open source development in a wide-ranging interview on TechRadar."
← Back to Stories (view on slashdot.org)
I don't know why, but whenever I try out KDE (every few years or so) there's something about it which drives me back to GNOME again. I'm trying out KDE4, which I really like but the real problem is program integration. The majority of useful utilities on Linux are written with GTK widgets rather than Qt, and while the Gnome-Qt bridge thingy which replaces GTK objects works for the most part, the different File Open/Save dialogue boxes grate on me.
I bought a laptop for my wife some time ago. It came with a extremely bad distro that tried to resemble Windows XP in every way - and, of course, it failed.
So, I asked her if she wanted to try Ubuntu, and installed it in the laptop. I had some problems with drivers for the webcam (which still doesn't work) and the wireless driver (which works using ndiswrapper).
She never typed a single 'apt-get' in the command line (in fact, I think she doesn't even know there is one) but, after the initial setup, I didn't have to help her at all. And now, even being an average computer user, she is trying to spread Linux to her friends and colleagues.
Any life is made up of a single moment, the moment in which a man finds out, once and for all, who he is.
However, it will be the compatibility with M$ software that will push Linux mainstream.
I wouldn't be allowed to install Linux on my laptop if didn't work with the corporate network. I would be fired if I didn't open the Microsoft Word documents.
I'm glad that Linux can talk Microsoftish, so I can use Linux at my will.
Exactly.
Microsoft compatibility is a misleading term as well. It is "Linux Compatibility With Microsoft Products."
Microsoft is not compatible with anything else but rather new microsoft products by design.
Linux, by design, is compatible with almost anything that it needs to interact with on a network. While this causes there to be proprietary code in your distro in some cases, it opens the door to actually using Linux.
The thing is, even if you are able to convert your whole company to Linux, companies do not operate in isolation, and hence the rep with his Linux empowered laptop will eventually have to go and do a presentation at a Windows shop, and hence compatibility will enable you to make the move to Linux with confidence.
Seven Days with Ubuntu Unity
Samba is a transition tool that allows businesses to move away from Microsoft. It's the leverage that a company's IT professionals (who are, more often than not, Linux-friendly) can use to transition away from Microsoft tools. When a company's Microsoft boxes are all talking to shares on Linux servers anyway, and saving scads of money doing it, it's more compelling to say let's begin to transition other things away from Microsoft.
Linux isn't going to displace Microsoft by ignoring it. Linux will displace Microsoft by offering reasonable alternatives at a more reasonable price, and by making the transition as easy as possible. To that end,
Samba is one of the best things that has ever happened to Linux.
Gnome is a problem because it doesn't encourage desktop use of linux in the sense that users use applications that are written to work and play well together. Gnome does nothing to encourage that.
This is the reason why OSX (and even windows) is beating the crap out of linux on the desktop and why it always will.
The average Gnome user (according to /.'s own roblimo) patches together a bunch of applications that are pretty much desktop agnostic.
The other two cases are the "optimizer" that worries about the terminal program they're using taking up another half kilobyte of memory when they recompile their kernel, or a developer that doesn't care about the desktop because all they really do is development.
All of these type of users would be just as well using Windowmaker or FVWM or any other window manager that we used to fight over last decade.
KDE has its issues, but at the very least it attempts to encourage users to utilize the K* applications and those K* applications actually work together very well. Further, the environment is a fairly consistent development target for applications.
At the moment, KDE is really the only coherent desktop environment that the free unix world has (with a possible nod to Xfce). Enlightenment development stalls out on a regular basis and as it stands now is years behind even Gnome for the most basic stuff. GNUstep was a great idea, but its in the same boat as E, too little, too late.
Finally, let's remember why Gnome was initially developed -- as a political response to some issues with QT that no longer exist -- and more often than not Gnome is still chosen over KDE in distros due to politics.
Gnome used to be my favorite window manager, until I realized that the focus model is totally wrong, and there is no way to fix it. It lies in this principle,
No window should ever steal focus ... ever. Don't even do it if the computer will blow up, or the world will end, if I fail to respond.
If you are trying to maximize your computing efficiency, focus stealing is your worst enemy. It's entire purpose is to slow you down.
Focus stealing breaks flow and can also be a security issue. Let's say you are typing away in a text editor and some other application decides it needs you to make a decision in the form of a pop-up window. Just as you are hitting enter in your text editor, the window pops-up, grabs your return keystroke and accepts it as your answer, doing whatever the default action is. Or the case where you are typing a password and a dialog steals focus and you type part of your password out in plain sight into the dialog.
A lot of this can be blamed on Windows for having a really shitty focus model, which everyone else tries to emulate in order to appeal to mouse-driven Windows users, I guess. I have noticed that Vista has a slightly improved focus over XP, but still very wrong. This focus model is the same attitude that gets you the Windows update manager that bugs you every 10 minutes by stealing focus, or worse, automatically rebooting while you aren't even at the machine (this is simply unforgivable).
Unfortunately, if you switch to a reasonable focus model, you will break poorly designed applications that are used to the broken focus models (OpenOffice, Matlab, any IDE, to name a few). These are applications that use a lot of pop-ups that don't "disable" the main window, which is when it is ok. For example, save dialogs are just fine for pop-ups: you can't have the main window in focus, so the change in focus to the dialog is natural and doesn't have a negative effect (it's not actually focus stealing).
However, pop-up text-searching is always absolutely wrong, for reasons beyond focus stealing too*. You will find that removing focus stealing will (correctly) not give these search boxes focus, which really breaks things for these applications. (Firefox wins here, maintaining its fairly good usability, with the integrated, incremental search bar.)
KDE does actually have a setting that can strictly stop focus stealing, in the form of a sliding bar. This works most of the time, but it's not perfect. New windows do in fact steal focus for a few milliseconds. This is enough to occasionally steal a keystroke, but I can live with it for now.
At work, I only get to choose between Gnome and KDE, so some other window manager out there may get this 100% right and I haven't explored it. At home I use IceWM, which also has a broken, unfixable focus model. However, the software I use at home is better behaved, making it less of an issue.
* Side rant here. For seasoned Emacs users, the incremental search function is frequently used for navigation (see item 4). If I need to move the point by more than a few characters or words, I start a text search (C-s, C-r) and type some text at the point I want to go. I do this all the time without even thinking about it. This doesn't work in a pop-up text search, even ignoring the fact that the aren't incremental either. When you bring it up, usually ctrl+f, there is always a delay to the window coming up. If I start typing my search right away, as I am used to in Emacs, it will go into my document rather than the search box. I find this incredibly annoying. I shouldn't be waiting for the computer like that.
??
1- NFS performance is amazing. It isn't the protocol you have performance problems with it is the transport (layer 1 or layer 2). The protocol in a transport might make a couple % points difference, and that even rarely.
The transport is where it is at. Comparing gigabit with FC is a losing battle for NFS, but compare 10G with FC (even 8G FC) and you have NFS at the top of the performance heap right now for mass storage, only iSCSI is in the same ballpark- but it is also on... 10G ethernet. iSCSI also cannot do simultaneous reads/writes like POSIX compliant NFS can. Direct attach is miserable because you invest loads in disk and can only use it on one server. What if you want to share that data around? Replicate? Islands of storage?
2) use automounter. Seriously, this hasn't been a problem for 5-10 years. Automounter, hostnames, don't use IP addresses (better if you can reverse the addresses).
You obviously haven't maintained NFS either recently or in a large environment.
NFSv4 does things your post doesn't even mention (security and ACL improvements, some performance in some cases).