A few notes that I scribbled while listening to this week's show:
1) Ubuntu's QML desktop - Fab, you wanker! Yes, Canonical is flailing but you keep attacking them for the wrong reasons. Did you even read the Mir justification on the Wiki? You can say that you don't agree with some of these decisions (like making the shell a different beast from other clients, in the name of "security"), but you attack them with all the wrong justifications: a) they did not say it will be ready for October on your desktop - they said it will be ready on the phone and if you'd have read the wiki you would have known that the phone and desktop use a slightly different stack (due to driver constraints mostly) and the desktop stack will be ready in 2014 - this is plenty of time for a team of full time employees (unlike Wayland which is a "spare time" project) read my blog post for my full opinion. b) even in 2014 they don't need to port the entire Ubuntu desktop to QML - Unity will be written in QML (which I've tried and its 100 times easier to write QML applications than GTK+ ones, and they already have most of the code around - "unity-2d" was written in Qt), and c) as Jezra noted - theming will take care of the look and feel - I'm running Kubuntu and all the GTK+ applications I'm using look just fine due to the QT-GTK+ theme engine.
2) Linus's secure boot apprehensions - what people keep ignoring for some reason (at least the pundits I read/listen to) is that Microsoft does not have a CA infrastructure and no one is expecting the Linux community to come up with one. What Microsoft has is a signing service, and the CA infrastructure is delegated to Verisign (in exchange for basically all the money that you pay Microsoft for signing - the $100 price tag for signing your boot module goes directly to Verisign). In order to provide a similar service for the Linux community without "deep throating Microsoft" as Linus succinctly put it, all that is needed is for RedHat (and SuSE and whoever) to come up with their own standard packaging and signing format and write a utility that takes a Verisign certificate and signs the module, then submits it to RedHat's signing service that will sign it with RedHat's secure key (also from Verisign, but this one is installed in all secure boot mobos) and you're done. This is exactly the process that Microsoft has. Setting this up should be easy for said $1bn-in-revenues company, and no CA and indemnification and $50mil/year required.
3) Firefox 3rd party cookies - I have to side with the industry official on this one: I've worked in the past in a company that does SaaS service for web publishers (not advertisers, and I'm not affiliated with them any more and have no vested interest in them - they shafted me on the stock options) and had to deal with Safari that had 3rd party cookies disabled since forever. It is really hard doing good content matching when you can't keep anonymous information on the user. "Yes, yes", you say to yourself, "anonymous my ass". And you'd be right, except that there is an official W3C standard for dealing with the privacy issues of 3rd party cookies - its called P3P and its a technical protocol to identify what information is being tracked by 3rd party cookies and what it can be used for, communicate that to the user and let them easily choose, and you know who have implemented it in their browser? Mozilla? nope; Google? nope; Opera? nope; Microsoft? why, yes they did - Microsoft is the only browser manufacturer that supports P3P and apparently actually cares about user's privacy and right to know what is being done with their information (and the ability to set security policies regarding that). Mozilla are wankers - instead of implementing an agreed upon standard, they use a big hammer - mostly because they have it lying around and they don't need to invest time in developing it. This will mostly hurt the web as there are many SaaS services that web developers use that will now loose their functionality for ~30% of the users (Firefox + Safari).
Regarding the Ubuntu story, I'm afraid you didn't get all your facts straight, Fab.
The Ubuntu installation is running on the phone's hardware without rebooting and without shutting down the Android stack. See here: http://m.youtube.com/watch?desktop_uri= … bunTRzEQCI .
Its done like a chroot basically. The Ubuntu setup is basically a live CD that is stored as an image on the phone's SD card, and because the underlying OS is Linux then you can just mount the image and chroot into it. You can also run X and get the full Ubuntu experience while the Android stack keeps the phone's standard UI and other facilities running.
What people have are doing with it now(as demonstrated in the video) is to run a VNC X server and connect to the VNC screen with a client, usually running on the same device. With Ubuntu's dock, you'd get the X output directly to your screen through the dock hardware.
Because the phone software keeps running, then you can get and receive phone calls and communicate with all if the other Android services. You can also run Android applications that you have on your phone, on the Ubuntu desktop. I'm not sure how that is done - there is a dalvik machine for X (no need to run an emulator) but I think you'd want the Android app to access the phone data and I don't see how they can do that when the use another dalvik VM.
Regarding the Motorola lap dock that you mentioned - that's a completely different (and much less awesome) approach - the simply turn off the main screen on the phone and start an android screen on the lap dock - its the same OS and capabilities, just with a larger screen.
Regarding the O2 phone number HTTP header debacle - I really don't understand what the fuss is about.
This use case is very common and actually required - for example, I'm developing an application for use by mobile phone users, and the application communicates with the user over both HTTP and SMS. The SMS section obviously requires the phone number of the user, and when the user then connects over HTTP, I need some way to identify that its the same user. Normally this is done by assigning the person a password and they need to type in their phone number and password (or get a link in an SMS that already encodes their phone number and they just need to type the password) - but the cellular provider that ordered this application wants that their users will not need to type in a password and be authenticated on the basis of using their phone on the cellular network - so we setup a "login method" where if a request comes from the cellular provider's WAP gateway and has the correct header set, we can identify the user and log them in automatically.
Some general comments about the episode's release news:
- Fab's comments about upgrading to 5.7 not being possible is probably right only if he meant that its not possible to upgrade to 5.7 from RHEL/CentOS 4 - which I don't believe is what he meant. Upgrading from any 5.x version to any later 5.x version is very possible and in my company we do it all the time. We just now completed an upgrade to CentOS 5.6 on servers installed using many different versions, some as old as 5.2 - no problems whatsoever.
- VirtualBox 4.1: Indeed the USB support is only available in the proprietary product, but as of VirtualBox 4.0.0, and unlike in 3.x, all the proprietary parts are in an "extension pack", and the version you can download from the web site is 100% identical to the version you get out of the box in Ubuntu, Fedora and the likes. As a user of the so-called "Open Source Edition" of VirtualBox 4.x you are granted free access to the proprietary extension pack under a "personal use only" license - so if you are not an enterprise, you can install the VirtualBox that comes with your operating system, then download the extension pack from VirtualBox.org and install it using the VirtualBox UI - and you get USB support completely legally. I won't get into why its done in such a way and I agree that its retarded, but USB support is available for non-enterprisy users.
I'm probably jumping too late here (I have a slight delay in episodes), but I wanted to mention something that somebody else has probably mentioned already regardin BTRFS and Fab's comment about lack of usefulness to users:
BTRFS allows snapshotting the file system and to create "recovery points" for installation - AFAIK Fedora 15 includes a yum plugin that uses this to "foolproof" installations so that if it breaks you can just revert to previous snapshot. I would have been pretty excited to have that in my Fedora 15 installation (which I upgraded from rawhide and was installed on Ext4) that I trashed when I did an install from updates-testing today.
My main problem with Unity is that it has no keyboard support at all - you can't even navigate the menus with a keyboard! Not to mention there is no "run dialog" that you can invoke to quickly run a program without browsing the menus.
A related problem is that you can't actually run a program if it doesn't have a global menu item under Unity - for example I have a couple locally installled application (installed in my user directory) which I launch using a shortcut from the panel - and you can't have that anymore.
Visually, OTOH, its very appealing. The unification of the window title, top panel and application menu (though very mac-like) looks very nice and saves a lot of space in a "vertically challenged" netbook screen. The only problem with that is that only pure GNOME application work that way and all other programs - including Firefox - now have two menu bars. Also the task panel on the left side is very nice looking and I like the "urgent" window hint effect that they did.
Finally, I didn't see any performance problem with Unity - it definitely performs just as well as GNOME Shell, better even in some scenarios.
And that was my 2cents about Unity
Nothing prevents you from writing other sync methods though...
That might not be possible without rewriting all the core applications, like contacts. Anyway, Android is a bad choice for people who don't like the corporate cloud, and nobody ever mentions that.
actually you can sync to an Exchange server (and compatible servers such as the Zarafa, the new Linux exchange replacement), and - not sure about it but i'll check when I get my new Milestone tomorow - you should be able to sync to any service that supports SyncML - like Funambol, or roll your own using any of the thousands SyncML implementations . i really like Mooha.
I don't think we have to worry much about the new upstart system. Several have tried to get it working, but it's so far from being close to anything useful that it's not even on the road map of being implemented even on experimental systems yet.
You mean, "the new init system", of course ;-)
After reading a bit more about this System-D stuff (including the origins of the name... very French), it seems like a good idea - its basically what upstart was set out to do, but make initialization order even more lazy: really really really only start services when they are needed, where "needed" is not defined as "I think service A uses service B, so please make sure it is started before you start B" (like in upstart) but instead "needed" is defined as - service A actually went and tried to open the communication port of service B.
Obviously there are a lot of obvious and not-so-obvious pitfalls that can crop up in the implementation, and I would hope that Poettering wouldn't take the approach of "developers should 'fix' their applications that don't conform to 'the new way of doing things'" and think a bit more about backward compatibility - and I understand everyone's concern that the guy has a track record of the reverse :-)
I guess time will tell.
MeeGo: works very well on a first generation AAO. ... The design is good and the UI is snappy, but one thing irritates me: the top window bar doesn't fit the design. Why did developers put a really big grey-silver bar, instead of something a bit smaller and with a colour scheme similar to the boarders?
I feel Meego still misses a lot of features that I came to expect from a windowing UI - like "maximize/restore". Otherwise its pretty nice.
You misunderstood me. Another launch system might be cool, I was doubting that Poettering is the right guy to do it, though. 1) He still needs to fix pulse and 2) I am highly sceptical of his ability to solve problems. It seems to me he tends to bite off more than he can chew and blames others. A launch daemon with the same kind of widespread problems as PA would be a nightmare...
Its not like I enjoy being contrary, but I for one really like PulseAudio - its not that it doesn't have its problems (such as RTP multicast flooding), but it allows me to hot plug devices such as my bluetooth headset or a USB audio port and easily configure everything - ALSA doesn't do that. Also, software mixing is good.
A few notes:
* Meego: Dan seems to be under the impression that Nokia will support Meego on the N900 devices, while the Meego 1.0 press release made it very clear that there will be a "developers release" for N900 devices, but that would not get support from Nokia. Instead Nokia will continue to support and issue updates to the Maemo 5 release.
* Nero are makers of an MPEG4 derivative called NeroHD that they sell to DVD set-top box makers (among others). under US patent system, if said DVD box maker sells NeroHD equiped boxes without buying licensing rights from MPEGLA then they are infringing on the patents. If Nero could shoot down those patent, their own icensing deals with hardware vendors will become much more attractive.
Guys, regarding your question on why people are upset with Xen [Dan: 00:24:10], the main reason is that XenSource's offerings - even before being bought by Cytrix - were all MS-Windows bound. You run your hypervisor on a Linux host, and you can run Linux VMs under that, but all the managing software (now known as XenCenter: http://www.citrix.com/lang/English/lp/lp_1688615.asp ) is all MS-Windows only!
So people understandably want to move to a solution which is more committed to open source operating systems.
Fab mentioned complaints that Google's support for the Nexus One is not very responsive with turn around time of 2 days for e-mail support requests(!!)
Wow - I say that is a great turn around time for e-mail support! I had some support requests with Nokia USA, and if I get a reply with in a week I'm very happy - usually turn around time for support e-mails for the largets mobile phone provider in the world is around 10 days.
Also regarding the KDE review: Fab commented about the three configuration options that can be found under a KDE application's "settings" menu. I would tend to agree that it can be very confusing - Just look at how the Kontact settings menu looks to get a head ache.
Of the two "extra" options that you often get in KDE applications - shortcut configuration and tool bar configuration, I would definitely get rid of the shortcut configuration dialog, in favor of GNOME's way of doing this (which is brilliant IMHO). But I really like to be able to configure the tool bars to remove buttons I don't use and add buttons for other command that I use. I would rather have the extra noise in the menu then loose that feature.
And I prefer to use cmus because it does things how I like them, rythmbox is however easier to use than both of the others, because it just hides away advanced features.
Which is an acceptable way to handle complexity, the problem is if you hide the advanced features so well its hard to find them. I don't think that Rhythmbox ha that problem, I just think that its missing a few features that I use daily.
Basically this is different folks, different strokes, that is why they are different projects after all. I myself can't stand amarok's wanting to be iTunes thing, and it is slow and clunky
It is, I fully agree with you here.
but again, that is my opinion, and nothing else, don't come here and tell me that you can say the one or the other. Give me configuration through readable textfiles, so that I don't have to worry about it when I am focusing on using the application like you have in vim and emacs and I am happy, I don't get what your problem is. Apart from you being dogmatic about what de to use, I don't like any of the DE's but you don't see me run around touting awesome the same way now, do you?
I won't presume to tell you what to use - I'm not dogmatic at all and I'm not sure why you'd say that. I use both GNOME and KDE on a daily basis (I use 2 workstations on a daily basis, where one runs Ubuntu and another either Kubuntu or Fedora with KDE). I only find fault in people saying things like "product A is bad because it doesn't look and behave like product B".
If you'd go back and see my previous comments in the Linux Outlaws forums, you'd notice that this is my theme - everyone should use what they like and they should not disparage other software projects on the basis that they have a different approach.
I just got to the KDE review section (I've been listening to the podcast in parts) and I must take offence with Dan's comment at about 1:06:30:
"KDE are not afraid to have 15 buttons to do 1 thing and GNOME they try to make less buttons do more functions."
While on the face of it this might seem a bit KDE friendly - it sounds like it only if you're a rabid GNOME fanatic . KDE are not afraid to let the user configure 15 different things while in GNOME its a sweet day in heaven if they let you configure 1 additional thing.
Compare for example Kopete to either Empathy or Pidgin (I'm not biased - choose whatever you think sucks less, even though Pidgin is not really part of the GNOME project):
- Which of these lets you turn off the people photos?
- Which of these lets you sort the contact list or group list automatically according to name or status? :geek:
And don't get me started on Amarok vs. Rhythmbox or whatever. If you don't like to be able to configure stuff, then use GNOME. I personally ( :ugeek: ) love the ability to change the layout of playlist items in Amarok because I wanted the playlist to show me how long it was since I last played that song. Let's see you setup Banshee to do that
And if they're video professionals, they probably wouldn't have a hard time figuring out how to import video with one program and edit it with another. I think we're talking about amateur video here.
I agree with mg6 here - DV import is not important for amateur level video editing and if you are a professional you probably won't have an issue with using a separate import tool (hey - its the UNIX philosophy - have your tool do one thing and do it well and use several tools to get a multistep job done).
Obviously I wouldn't have minded a good video editor that also does DV import (I actually have an amatur camcorder that only does DV output, its just that i haven't turned it on in over 3 years ;-) ), but given the choise of putting work on DV import or better video editing, I'll choose the latter every time.
i like the idea of $1 US coins. they would last longer as well. no need to keep printing paper 1's all the time. the problem is, they usually don't look that great. the current ones are about the size of a quarter (psychologically that may be a hindrance to their acceptance). and the gold color tarnished and looked crap.
the US should make some nice looking $1 coins. the 1 pound coin is quite nice. its thick and heavy. it feels significant. and they have a neat 2 pound coin. large with in inset. a US $1 coin should "feel" like more money then a quarter.
The Israeli 1 NIS coin is actually the smallest coin in circulation (and probably the smallest coin ever created in Israel) - its about the size of a dime although a tiny bit smaller in circumference and heavier. People really love to use them, because you can fit a lot of them in a small pocket. The new 2 NIS coin (called lovingly "shnekel", which is a portmanteau of "Shney" meaning two and "Shekel" which is the name of the currency) is also rather small and is even more popular.
One last thing to set straight though - regarding the show's main feature of the N900 review:
Both Dan and Fab have commented that Nokia has plans for only one more Maemo device, after which they will abandon the platform (Episode 128, starting at 1:38:00). I'm not sure where they got that from. Nokia did confirm another Maemo device for the second half of 2010, based on the newer Maemo 6), but they also said that by 2012 all N series phones will be Maemo based - this doesn't sound to me like dropping the platform.
Also, a friend of mine got himself a N900 and the screen is very solid - completely unlike the touch screen on my Sony Ericsson. I think it is resistive and not capacitive because you can use it with the plastic stylus, but it feels like solid glass - very like an iPhone screen and not like the old touch screens that feel a bit "squishy". As Dan noted, the N900 screen is very fingerprint friendly (not like the new iPhone 3Gs screens that are rather resistive to this stuff) and so it comes with a cleaning cloth (very weird, if you ask me) and it also not very scratch proof - as my friend unfortunately found out :whistle: , more like 1st gen iPhones really.
I just looked through all your posts, and I see that you haven't posted anything else than criticisms on the show, and most of the time you take things out of context or redefine the meaning of things being said in the show, so it is not so strange that he gets a bit annoyed...
Yep, I'm sorry about that - I try to be a positive member of society, but frankly I fail a bit at that, at least when Linux Outlaws is regarded. I can blame the fact that I listen to the show while driving (where I obviously don't have access to a computer) so I mostly comment about things that I think the guys got totally wrong and I feel very strongly that I have to set the record straight.
I'll try to be better :oops:
Regarding the tray icons - the reason that there are very few applications (other then the applets that ship with the desktop environment) that have tray icons, is that writing tray icons in Linux is fscking hard!
Once again, I never said that. I said they are all configurable and in general Linux apps are more considerate about having a tray icon. Please learn to listen to what I am actually saying.
You said, on Episode 12 starting at 0:56:55 - "that whole advertising thing [where every app has an icon in the tray] that's not happening in Linux. I actually think that Linux applications are really careful about how to put an icon in there". Well - they are not careful, just that most application won't go the whole way to put icons in the system tray because its very hard. If Linux (or actually freedesktop.org) had a good protocol for notifications icons (which they do now) and there were good library implementations for that (still working on that part), then you'd start seeing every single app showing icons in the tray.
BTW - the system tray is called "Notification Area" in GNOME and the GNOME Shell (coming in 3.0) changes that to be something completely different (so is change good? is change bad? is change good only if it comes from upstream?).
This is I think the third time you responded with "I didn't say that" to comments I made about the show. I really hate that - I comment on things that I hear. If I referenced to things out of context then I apologise in advance, but I never invent stuff that you said. From now on I'm going to add time code references to every comment I make - I hope that will mean we can drop the "I didn't say that" shtick.
Time codes may not be accurate - my podcatcher tends to mess a bit with the time stamps when I rewind.