Of the three books I looked at, Chobits was the closest I came to actually buying.  In Borders it is $24.99, on Amazon it is $17.99.  I have a hard time overlooking $7, especially when the brick and mortar is going to add tax, which would make the different over $9.  I wasn’t sure if I wanted the book, which is what actually stopped me from buying that particular item.  The next two books… were much more clear cut.

Next on the list is The Official Ubuntu Server Book.  In Borders it is $39.99, but only $26.39 on Amazon.  A $13 difference (Plus Tax) is just not even close.  After flipping through this book, I decided it was actually quite nice and rather wanted to buy it.  So I purchased it from Amazon right there in Borders, using my Amazon Android application. The book will be here shortly.

Last was the LPI Linux Certification in a Nutshell book.  In Borders it is a tech-book typical $49.99.  Demonstrating that it just isn’t financially feasible to use brick and mortar stores is Amazon with a price tag of just $31.49 which is an $18.50 difference (not including tax).  I realize that the $49.99 value is list price, but how can any consumer seriously consider spending 60% more on one single book, just to buy it in store.  Sure, if you absolutely, positively MUST have the book right that second, you’ve got no choice.  If I had any option whatsoever, I’d Amazon it.  $3.99 for overnight shipping is still much, much cheaper.

Yes, I like to be able to pick up the books and flip through them before I buy them.  Yes, I like to just browse the shelves of a bookstore.  But no, I will not pay a $39.10 premium to do so.  That is enough to go and buy a copy of Ubuntu: Up and Running: A Power User’s Desktop Guide AND the Ubuntu Pocket Reference Guide AND still have $6.00 left to buy a coffee at Starbucks.  Sorry, Borders loses.

This proved not to be so easy as the Roku didn’t seem to fit into any of the existing categories of Television, Amplifier, Cable/Satellite Box, Video Recorder, DVD, Music Player, Game Console, Mini System, Computer, Home Automation, or More Device Types. Well, I thought it fit into Mini System or Home Device Types, but those two categories were comprised of Mini System (DVD, CD, Radio), Mini System (CD, Radio, Cassette), and Mini System (DVD, VCR, Radio), and Tape Deck, Minidisc Player, and DAT, respectively. After some playing around I finally found it. Apparently, the Roku is a DVD Player! Or at least that’s what Logitech says.

Inside the Logitech Harmony Remote Software (I have version 7.7.0), do the following:

1. Go to Devices
2. Click Add Device
3. From the first dropdown, select DVD as the category and DVD as the subcategory
4. Select Roku as the Manufacturer
5. Enter your model (mine is an N1100, the example is N1000)
6. Click Next
7. It will show your device, then click Next again
8. Scroll through your list of devices and click Settings on the one named DVD (this is your Roku’s default name)
9. Select Rename this device from the list
10. Click Next
11. Enter the new name (i.e. “Roku” or “Roku HD”)
12. Click Save
13. Click Done

Congrats! You now have your Roku player associated with your Logitech Remote. Now you can add it to any activities you like or create new ones.

The Screen – I never really thought of the screen on the Kindle 2 as “pale”, but looking at the K2 and K3 side by side, text on the K3 is much darker.  Amazon says the screen contrast has been up’ed by 50%, and I’ve got no way to prove that – but it does look much nicer.  They also made the screen flips faster.  I’m not sure how much faster, maybe another 50% faster.  Needless to say it is very fast and shouldn’t be problem unless you’re stabbing the page button as fast as you can.

Size and Weight – Amazon says the K3 is 21% smaller and 15% lighter than the K2.  This is obvious right off the bat; as soon as you pick up the device you will feel the difference. At this point, the weight of the device only makes a difference for long reading sessions, but I know the K3 will be nicer to hold in the longer run.  As for the size, if you take off the top bar and right bar of the K2, you’ve got the size of the K3.  The one thing I will note, the K2 and K3 are exactly the same thickness, so nothing has changed there.

The Exterior – Everyone knows the Kindle 1 was not exactly a sleek device.  The K2 definitely fixed that issue.  The K3 improves yet again upon the sleekness.  First off, they moved the speakers to the top (back) of the K3, so your hands aren’t covering it.  Second, they moved the power button, headphone jack, and volume control to the bottom of the device.  I do mean the absolute bottom, on the same line as the power charger (which is still Micro-USB, thankfully).  Lastly, One very nice and unexpected change was the case itself.  They made the entire unit one type of plastic (no more brushed metal back).  This (what I am assuming is a new) material also has been given a slight rubbery feel.  It isn’t so much that it feels like you are holding a tire, but just enough so the Kindle 3 doesn’t feel so slippery.  Plus you can prop it up on a slippery surface (like a hard plastic palm rest of a keyboard) and it is much less likely to slide.

The Speakers – While the Text-to-Speech feature on the Kindle is nifty, I rarely use it.  Generally, when I want to read, I want to actually read.  That being said, I’ve been known to load up a few MP3s on my Kindle 2 so I can listen to some tunes as I read.  Just for fun, I loaded up the same MP3 on the K2 and the K3 and I’ve got to say, the Kindle 3 sounds much better. The Kindle isn’t going to win any awards as an MP3 player, but the speakers do play a fairly good rendition of the music.  They are much less prone to clipping on the high volumes, and possess a much more balanced tonal range.  Heck, I can even feel the slightest hint of bass coming from the unit on the drum beats.  Sure, it isn’t going to blow out any windows, but for an eBook reader, it is impressive.

The Page Buttons – So the Kindle 1 had HUGE buttons that were way to easy to accidentally press.  In the refined Kindle 2, the buttons were much smaller, but still had a tendency for accidental activation.  For me they were a problem after longer reading sessions where I ended up wanting to hold the K2 almost on top of the button because that was most comfortable for me.  In the K3, the buttons are tiny.  The K2′s buttons measure 1/2″ in width, the K3′s measure in at about 3/16′s of an inch (that’s less than half the size).  Effectively, the K3′s buttons are just the edge of the unit.  They’ve also changed up the layout a bit so Next Page and Previous Page are on both sides, while Home, Menu, Back, and the direction pad have been moved to keyboard.

The Keyboard – With the exception of the addition of the Menu, Home, and Back keys, the keyboard doesn’t look much different.  They did remove all the numbers (they now appear after clicking the Symbols button) and the forward slash.  With the shrinkage of the case, the keyboard also takes up less real estate.  The keys themselves are the same size, but the space in between them is greatly reduced.  I haven’t done much typing yet, but the keyboard seems to be much of the same.  The new direction pad, on the other hand, is quite nice.  It is smooth and integrated.  Most importantly, the K3′s direction pad doesn’t stick out like the K2′s joystick.

The Reading Interface – Most of the Kindle 3′s interface is just like it is on the Kindle 2 (nothing major enough to worth noting), but they made a number of changes to the reading interface which are quite nice.  First off, once you are into a book, the title bar (containing title, time, etc) disappears.  If you press the menu button, it reappears to tell you the book title, time, wifi status, 3g status, and battery status. The only difference from the K2 is the fact that the K3 no longer shows you the amount of free space.  Note: this is only when reading a book; when you are at the home screen, pressing Menu will still show you the free space.  The second change was to get rid of the bottom “bar” of empty space that you had in the K2.  In the K3, the location bar is now directly flush against the bottom of the screen.  These two changes combine to give you 3-4 more lines of reading space out of the same screen (on the smaller font settings).  The third change is the addition of new options in the font settings.  In the Kindle 2 you could change only the font size and “Words per line”.  In the Kindle 3 they have also added “Typeface” which gives you the options of “regular”, “condensed” and “sans serif”.  You also have “Line Spacing” with the options of “small”, “medium”, and “large” which reduces the amount of vertical space between lines (the “large” option is the default, and the size the Kindle 2 used).

In conclusion: If you don’t already have a Kindle and have been thinking about it, go buy a Kindle 3.
As for whether it is worth the upgrade, to me it definitely is, heck I thought it was worth it when I first heard about them, which is why mine arrived yesterday (Editor’s note: Jon preordered THE moment he heard about it.).
The question you should be asking yourself is: Are there any items in this list that I really wish I had on my Kindle?

I had read somewhere that “Map Network Drive” was supposed to work with WebDAV in much the same way as in OSX.  I tried it out, but it never seemed to work.  I spent quite a while googling about for instructions that would work and eventually landed upon a semi-promising article that mentioned the need to install the “Software Update for Web Folders (KB907306)“.  Cool! Maybe that was all I was missing. I gave the update a shot, and a reboot later I was ready to WebDAV…

Of course it didn’t work.  That update does allow you to put in URLs directly into “Map Network Drive” rather than having to go into the more hidden “Connect to a website…” option.  Even still, I couldn’t connect to my WebDAV share. I saw the connection attempts on the Apache logs, but Windows kept telling me it wasn’t a valid share (Liars!).  After some more research, I found a post that said that the WebDAV connector didn’t allow ‘basic’ authentication, but you could fix that with a registry hack.  This time I knew better than to get my hopes up… which was a good thing, because it didn’t work either.

After still more research, I found the answer.  Microsoft’s WebDAV client only supports the use of ‘digest’ authentication.  I almost never use digest authentication, even though it is technically superior, simply because it is not well supported.  Any time you need “secure” authentication, you use SSL.  In the interest of being sure, I switched my authentication on the WebDAV share from ‘basic’ to ‘digest’… TADA… Windows 7 worked with WebDAV.

Of course, there is one CRITICAL flaw, I’m doing this against an LDAP back-end.  You cannot mix ‘digest’ authentication and LDAP.  The short reason is that because digest hashes the password before sending it (where basic sends the password clear text), and that hash is not compatible with your LDAP password hashes. It makes sense, and is nether LDAP’s nor Apache’s fault for not working.  The reality is that it is dumb (and typical) of Microsoft.

In summary: You cannot have Windows 7 natively talk to a ‘basic’ authentication WebDAV share, Period.  You can use Windows 7 native WebDAV with ‘digest’ authentication only, but that prevents you from using LDAP.

To connect to a WebDAV share under OSX:

• Finder
• Go
• Connect to Server
• Enter your URL – The same as you’d use in a browser (IE: http://fileserver.company.com/webdav/ )
• Connect
• Enter your Username & Password (provided it is required – just like the web browser)
• Connect – That’s it!

To connect to a WebDAV share under Linux (Ubuntu 10.04):

• Places
• Connect to Server
• Service type: WebDAV
• Fill in Port and Folder
• Connect – That’s it!

Components used:

• Ubuntu 10.04
• Apache 2.2
• OpenDS 2.2 (LDAP)
• mod_authnz_ldap, mod_dav, mod_dav_fs, mod_dav_lock & mod_rewrite

The process is very easy, provided you’ve got OpenDS & Apache 2 already up and running. If you need the basics of that, I’ve covered that in another post. I’m also assuming that LDAP is already setup with users & groups.  In the following example, you’ll have 3 groups: “All” which contains a list of all users, “One” which only gets access to the ‘share’ folder of ‘one’, and “Two” which is the same concept as “One”. Note: I used OpenDS Static Groups, I have no idea if Dynamic Groups or Virtual Static Groups will work.

• mkdir /home/webdav; mkdir /home/webdav/one; mkdir /home/webdav/two; mkdir /home/webdav/public
• chown -R www-data.www-data /home/webdav
• chmod -R 755 /home/webdav
• chmod a-w /home/webdav
• a2enmod authnz_ldap dav dav_fs dav_lock rewrite
• nano /etc/apache2/sites-enabled/000-default
• Add the following to the bottom of the file before </VirtualHosts>

Alias /webdav/ "/home/webdav/"

<Directory /home/webdav>
 Options Indexes FollowSymLinks MultiViews
 AllowOverride AuthConfig
 Order allow,deny
 allow from all
</Directory>
DavLockDB /tmp/DavLock

RewriteEngine On
RewriteRule ^/webdav$ /webdav/ [R=301]

<Location /webdav>
 Dav On
 AuthName DAV
 AuthType Basic
 AuthBasicProvider ldap
 AuthzLDAPAuthoritative on
 AuthLDAPURL "ldap://127.0.0.1:389/ou=People,dc=DOMAIN,dc=TLD?uid?sub?(objectClass=*)" NONE
 AuthLDAPGroupAttributeIsDN on
 AuthLDAPBindDN cn=USERNAMEHERE
 AuthLDAPBindPassword PASSWORDHERE
 Require ldap-group cn=All,ou=Groups,dc=DOMAIN,dc=TLD
</Location>

<Location /webdav/one>
 AuthName DAV
 AuthType Basic
 AuthBasicProvider ldap
 AuthzLDAPAuthoritative on
 AuthLDAPURL "ldap://127.0.0.1:389/ou=People,dc=DOMAIN,dc=TLD?uid?sub?(objectClass=*)" NONE
 AuthLDAPGroupAttributeIsDN on
 AuthLDAPBindDN cn=USERNAMEHERE
 AuthLDAPBindPassword PASSWORDHERE
 Require ldap-group cn=One,ou=Groups,dc=DOMAIN,dc=TLD
</Location>

<Location /webdav/two>
 AuthName DAV
 AuthType Basic
 AuthBasicProvider ldap
 AuthzLDAPAuthoritative on
 AuthLDAPURL "ldap://127.0.0.1:389/ou=People,dc=DOMAIN,dc=TLD?uid?sub?(objectClass=*)" NONE
 AuthLDAPGroupAttributeIsDN on
 AuthLDAPBindDN cn=USERNAMEHERE
 AuthLDAPBindPassword PASSWORDHERE
 Require ldap-group cn=Two,ou=Groups,dc=DOMAIN,dc=TLD
</Location>

• /etc/init.d/apache2 restart

At this point (provided you changed DOMAIN, TLD, USERNAMEHERE, and PASSWORDHERE in the above example), you should be able to point a browser to http://yourserver/webdav and it will prompt you for your username and password.  To Apache /webdav and /webdav/ are different, but most users won’t know that, hence the redirect.  After you authenticate, provided you are in groups “One” and “Two” you should be able to see 3 folders (one, two, and public).  If you are not in all the groups, you will not see nor be able to access the folders (except public, which all authenticated users would be able to see and access).

If you’re using a Mac, you can use Finder > Go > Connect To Server with the same URL.  You should be able to simply drag and drop files on and off (like you would any other type of share).  All directories, that you’re a group member of, except the base (/webdav/ – That was the ‘chmod a-w’ line) should be writable.  You can pop open files and edit them directly from the webdav share too.

Piece of cake, eh? Without clear and concise instructions (such as above) it took me less than two hours to research, figure out, and implement.  If you can follow the instructions and have some idea what you are doing, you should be able to get WebDAV shares up and running in less than 30mn (and that’s on the outside).

NOTE: You cannot use digest authentication, you MUST use basic authentication.  This sends passwords in clear text.  If this is internet accessible – I highly recommend you SSL your WebDAV share.

If you stop by OpenDS’s website, you’ll find a big “Get 2.2 Now” button. This does a WebStart install, which I’ve never seen before, but works fairly well on Windows and OSX environments. I’m sure it would also work under Linux, provided I had a Linux machine that ran a GUI (which I generally don’t). So these instructions will cover installing it from the command line, and some basic post-install configuration from the GUI (it’s just so much easier).

• apt-get install python-software-properties
• add-apt-repository “deb http://archive.canonical.com/ lucid partner”
  • From Lucid Release Notes
• apt-get update
• apt-get install sun-java6-jre unzip
• wget http://www.opends.org/promoted-builds/2.2.0/OpenDS-2.2.0.zip
• unzip OpenDS-2.2.0.zip
• mv OpenDS-2.2.0 /opt/opends/
• /opt/opends/bin/create-rc-script –outputFile /etc/init.d/opends
• update-rc.d opends defaults
• /etc/init.d/opends start
• cd /opt/opends/
• ./setup
• Answer the on screen instructions, the defaults should be acceptable in most cases, but I prefer to turn on SSL and StartTLS

That’s it. Like I said, not exactly rocket science. If you do the WebStart on a machine with a GUI, you’ll be installed in even less time. After that you can use the GUI to populate your server instance with useful data. So on your GUI equipped machine (which you’ve installed OpenDS onto):

• Launch ./opends/bin/control-panel (Or ./opends/bat/control-panel.bat – for Windows)
• Select “Remote Server”
• Enter the IP address
• Leave port 4444 (unless you changed it)
• Change ‘cn=Directory Manager’ to your admin username
• Enter your password
• Ok!

At this point there are two important screens for someone who wants to get up and running pronto:

• Schema > Manage Schema – The allows you to see what schema is built into OpenDS (basically everything in the LDAP RFC’s) and lets you add your own schema.  Simply hit “New Attribute” or “New Object Class”
• Directory Data > Manage Entries – This is where you’ll go to add/delete/change all the information in your directory. I think this interface is probably the best of all the LDAP interfaces I’ve tried (Including Apache Directory Studio and phpLDAPadmin)

I would suggest creating (under the base DN you specified during the install, something like ‘dc=test,dc=com’): At least 2 new Organizational Units.  ‘ou=People’ and ‘ou=Groups’.  Those two are the standards for Users and Groups, respectively.  After you have the ou’s, you can right click and create new people or groups.   A word of advice on Groups, I haven’t had a lot of luck on application compatibility with “Dynamic Group” or “Virtual Static Group”.  I’ve found it is much safer to stick with “Static Group”.

We started off by researching and trying the LDAP providers. After some investigation and testing of OpenLDAP, we settled on OpenDS. I didn’t like OpenLDAP (though it is something of the “standard”), because it is just such a pain in the ass to administrate. Everything is command like and in LDAP-ese, which is probably ok if you’re a regular LDAPer, but I’m not and neither was the local IT department. OpenDS, while Java based (and I dislike Java on principal), was quite nice. You can click the 1-button “web start” and have it install and do everything automagically. I’ll (probably) cover it in more detail in a later post.

So now we have OpenDS 2.2 setup and running well with the prospect email system. Next on the list of “critical” services was Samba, which is what they were already using and Samba claims to be LDAP compatible. Since we were running Ubuntu 10.04, I simply apt-get installed Samba 3.4 series and the necessary LDAP tools. This is where things started to go poorly.

I won’t bore you with all the debug details, but I ended up spending probably 4 full days worth of time trying to get Samba working against LDAP properly. I’ll admit that it could have been slightly faster, as I’m not terribly experienced with LDAP and had to learn some of it on the go. That being said, I’m used to open source, and the requisite learning on the go. From the best I can tell, Samba 3.4 makes a number of small changes which are basically undocumented and is designed to talk to OpenLDAP (slapd) only. One of the “small” changes in Samba I noticed was that the smb.conf option of “valid users =” is now simply “users =”, and doesn’t notify you of it’s deprecation. It simply ignores “valid users =” as an unknown option. Fan-freakin-tastic.

In the end, I could not manage to get Samba to talk to OpenLDAP using SSL or StartTLS. I did eventually get Samba to authenticate using my LDAP accounts, but that required a LOT of special objectClasses and extra miscellaneous information in the LDAP directory. I don’t mind having to have extra information in LDAP, as that is what it is good for, but having its own style for group user lists, userids and password borders on asinine. I know it is a throw back to it’s original building, but UPDATE THE CODE PEOPLE.

In the end, Samba was abandoned because it flatly ignored my groups. Once you authenticated to Samba with an LDAP valid username and password, you had access to every share. I tried my damnedest to make shares in Samba that I couldn’t access. I set up special groups, I used groups that didn’t exist… everything I could think of and nothing worked. There is a point of diminishing returns and I had hit it. In the end, with a mostly Linux/Mac user environment, there is no reason to use Samba.

I found a better option, WebDAV. I configured Apache 2.2, with mod_dav and mod_authnz_ldap. In an hour and a half, I had it up, providing access to files from multiple “shares” AND authenticating properly to OpenDS’s LDAP (over SSL mind you) with full support of my existing groups. Want to know how many changes I had to make to my base LDAP schema (sans Samba junk) to get this working? None. Not a single change at all. Kick ass.

So to summarize: Samba is old and outdated. It shouldn’t be used to share files unless absolutely necessary. While supposedly supporting LDAP in general, it doesn’t work. Most of the Samba/LDAP configuration is centered around acting as a PDC (Editor’s note: Primary Domain Controller) and very little covering simply sharing files. Lastly, it’s documentation (the official Samba website) is old and broken, literally 404′ing in many cases.

Since the iPhone burst onto the scene in 2007 with its “game changing” multi-touching, nothing has been the same. There have also been numerous articles as of late on how mobile devices are outselling PCs. Some of them are fairly vague as to what exactly is a mobile device (i.e. do Laptops count as PCs or mobiles?). Irregardless, it does go to show that there are a LOT of mobiles selling, and that will continue to be a big market in the future.

Linux has been predominantly server-only as far as operating systems go. Sure, you can use Linux on your desktop, and I had as far back as Red Hat 6.0 in the 90s, but it wasn’t an end-user friendly experience. Canonical and Ubuntu have been slowly but surely changing this, with their “For humans” drive. They’ve even gone as far as releasing Ubuntu Netbook Remix (which I’ve covered a number of times here) which is rather slick as far as netbook OSes go.

If Ubuntu can get GUI software projects to include multi-touch support in their applications, the game for larger mobile devices, tablets and netbooks, could change radically. A true and proper operating system (that can do more than just watch videos and punch out email, like the iPad), with great touch support could potentially put a serious hurt on the iPad. Sure, part of the iPad’s appeal is the hardware, and no arguing that is nice (for what it is), but hardware is easy to replicate.

One of the biggest challenges for Linux adoption has been the fact that it isn’t Windows or Mac. Simply the lack of ability to run your favorite applications is the biggest drawback for most people. On the mobile/tablet device market, no one cares. Windows Mobile, Android, Blackberry, iOS… all of them are custom made for their devices (cell phones) and form factors, and no one cares. For Tablets you choices are only iOS (on the iPad) and Windows 7 (for everything else). While Win 7 for Tablets is the same OS, people don’t care nearly as much. The tasks you want to accomplish on a Tablet are different than those on a desktop or notebook.

As much as I’d like to hope for it, Ubuntu 10.10 (Maverik) isn’t going to drastically change the game. It will show us the potential of Open Source Multi-Touch, but in a limited fashion. After all, the release is only 2 months off. Ubuntu 11.04 (P) will probably refine muti-touch nicely and add support from a number of applications. One (specifically this OSS nerd) can only hope that the rest of the Linux community will adopt the same standards, if not the same code base. Fragmentation of this new “basic input” would be lethal.

Like a good Open Source nerd, I Google’d about and found a potential answer. You need to:

• Boot into Grub (hold down shift after the BIOS finishes, and before the Ubuntu logo comes up)
• Select your default boot option, hit “e” (For Edit)
• Change the launch line (which starts with ‘linux’) and append “forceversa i905.mode=0″.

Except in my case, that was already there. So I tried removing this. Now the computer booted (heard the Ubuntu startup noise), but the screen was still dead. I had caught a line somewhere that someone had a similar issue and that it was simply the machine redirecting all graphical output to the external display. I plugged in an external LCD and confirmed I was getting output.

Since I had a semi-functional machine at this point, I hoped that doing an apt-get update / apt-get dist-upgrade would solve the problem. The machine was a clean install, so it needed 250 some updates, including a new kernel. New kernels are always promising in my book. After waiting for several hours (slow internet), the update completed and I rebooted the machine. No dice. The latest Kernel for 10.04 Lucid, does nothing.

Eventually I found this bug entry, which I believe is the issue. The last comment, from just about a few days prior to my fix attempt, indicated this will be resolved in 10.04.2. This is great news, except for the fact that .2 isn’t scheduled to be out until late January 2011. But that at least gave me hope that there was a fix in the works.  I decided to take shot in the dark and tried installing the kernel daily build.  The kernel I installed was from 2010-07-27 and works.  By works I mean the machine boots normally AND the screen displays properly.

So, if you have this issue and don’t want to wait, use a daily build kernel.  Now, I really dislike using a non-tested kernel like this for a standard user’s machine (if it was for myself, it wouldn’t be a big deal), but there isn’t much of a choice.  Without the new kernel, there aren’t many other options, other than telling the user “Sorry, you can’t use this computer” – and that tends to go over poorly.  So, new kernel = good (fixed), but can add some risk.  Let’s just hope the user doesn’t dist-upgrade to a new (older) re-broken kernel.