Community Matters

Lately traffic on this blog has gone up quite a bit (more than quadrupled over the last couple of months). And the main driver of all these new visitors? Google.

More than 90% of the readers of this blog come from search now (up from less than 40% in December) and again more than 90% of the search hits come from Google.

Looking at the traffic and at the actual searches that got people to my blog brings some surprising results:

• My own blog is the 162nd hit when you are searching for dirk hohndel blog. WTF?
• When being less specific (searching for hohndel blog) it shows up as the fourth result (and in Google’s defense, two of the three higher ranked results are my blog at work)
• On the other hand this blog is the top result for queries like eeepc competitor or motorola ming leopard.

Which brings me back to the title of this post - even though this is clearly my blog (the posts are authored by “Dirk” and the domain is “Hohndel.org”) and even though I write about lots of things here, most of the people who get here through Google are not looking for me but for information on a small subsets of the topics that I post about. And a change in Google’s ranking (which I can’t really influence, can I?) will completely change the amount of people coming to this site and the topics that they might be interested in. Something to think about.

So besides silly search engine optimization tools - is there an easy way to fix the fact that searching for my name doesn’t get you here? Here’s one idea: if you happen to have a blog or site yourself… would you mind adding a link named “Dirk Hohndel’s blog” which points to the entry page of this blog (so something like <a href=”http://www.hohndel.org/communitymatters”>Dirk Hohndel’s blog</a>)? And drop me a line if you did (I’ll happily add a link here in this post back to you). I want to try these searches again in a few weeks and see if the rankings respond…

Thanks for visiting!
I hope this was helpful - if not, please leave a comment and let me know why! Were you searching for something else? Did I miss an important aspect?

The good folks over at engadget were talking to Sony Vice President Mike Abary - and in that conversation he apparently decried that devices like the EeePC would create a race to the bottom. Interesting perspective - but it may be slightly influenced by the fact that Sony doesn’t have a competing product and would rather not have to invest into a market segment that could seriously impact it’s ultra-light notebooks.

But instead of complaining about the lack of features (which seems to be implied in this comment), I think Sony should try what HP and Pioneer and others are trying: create a competitively priced product that provides better features and therefore can grab a higher price and maybe even higher margins.

I certainly think we are going to seeing many more products below $500 that don’t go for “cheapest” but for “best value proposition” instead. And for Intel, AMD and Via this is indeed a very interesting market. The margin in their business can be influenced two ways. Higher ASP (unlikely here as the OEMs can’t allow the BOM (bill of materials) to get too high) or lower production cost. That’s where tiny chips like Silverthorne, created on a 45nm process, come into play. If they can be produced sufficiently cheaper compared the current entry level CPUs (like the 65nm based Intel Celeron processor) then it is possible to have a similar or even better margin percentage, even if the resulting chip carries a lower sticker price. Via already announced their next generation Isaiah processor (still on 65nm, though) which will compete in this market, and I am sure that once AMD can produce its Geode processors on 45nm, they will see better margins from that, too.

I love a good story - but this one I’m kind of skeptical of… guryhwa even posts some pictures of upgrading his EeePC with an Intel Pentium M, but the part where he says “soldering is complicated but interesting” seems to be understating the problem a bit. And he doesn’t have a picture with the CPU removed, only a picture of what looks to be a Pentium M ULV 753 on the Asus motherboard.

Call me unfair, but I somewhat doubt that this is really what happened - the labeling on the CPU is easy enough to create with Photoshop, same for the BIOS screen; much easier than the soldering work that would be required to do this.

But maybe he’ll post a few more pictures of the process to convince skeptics like me? It would be the coolest EeePC hack, ever.

This seems to be a surprisingly frequent question - at least if I consider how often this or one of its permutation ends up being a search term than lands people on this blog.

So here’s the simple answer. When you set up you EeePC you are asked to set a password for your account. That same password is the root password. So for the Asus / Xandros OS there isn’t really a default password; you set it yourself.

If you are running EeeDora on your EeePC, then the install script sets a default password of eeedora. My recommendation would be to change that right away…

Ever since more and more small formfactor “cheap” laptops started showing up (it seems the segment really got started by the EeePC, but lately the Cloudbook, the HP Compaq 2133 and the Kohjinsha E8 have sparked my interest) I have been wondering about the relative performance of the underlying platform.

There are tons of sites analyzing the performance of the high end chips of AMD and Intel (Phenom and Core 2). From Anandtech to Tom’s Hardware and many other sites. But there’s very little data to be found on the lower end chips that I mention here. Imagine my excitement when I found one at Tolly Group - but then I saw that the study was commissioned by Intel; that makes it a little harder to take it at face value…

But even after a lot of googling, this is the only benchmark based comparison that I was able to find. And given the CPUs that it tests (the same one as in the E8, a faster one than in the Cloudbook and a slower one than in the EeePC), it does seem at least somewhat fair to draw conclusions from it.

The summary of the the results is easy - the Intel Celeron beat the AMD Geode and Via C7 in all benchmarks: SPEC CPU 2000, PCMark 05 and SANDRA 2007. Again, take it with a grain of salt, given where this comes from, but it does seem to confirm what I have read so far about the perceived performance of these systems.

Oh, and just for the record - even running at the full 900 Mhz, the EeePC feels a bit slow. Hopefully the Gen2 system will use a newer Intel platform (Diamondville?) and be a bit faster. But unsurprisingly, there’s no publicly available performance data for that, yet.

I am a Lightroom user - have been since their very first public beta release. But this latest announcement from Apple has me interested in Aperture for the first time… tethered shooting, integrated into my digital asset management system? That would be great.

Apple says that they support only a few Nikon and Canon DSLRs. The list includes the Nikon D200, D300, and D3 as well as the D40x, D70, D80, D2Hs and D2Xs (that seems like a good list). On the Canon side they mention only the Rebel (not Rebel XT or XTi), 5D, 10D and 1Ds MkII - so basically all the recent Canon cameras are missing.

Let’s be honest, this is not enough to get me to migrate from Lightroom to Aperture - but maybe Adobe will integrate similar functionality into Lightroom?

Competition is good.

Wow, they just keep coming in. This is the second new competitor for the EeePC that I found today. The Kohjinsha E8 (the name doesn’t quite roll off my tongue, yet) allegedly will start shipping next week.

Just like the HP I blogged about earlier today it will be more expensive - around $600. It’s screen is small (7″) but at least has a better resolution than the EeePC (1024×600) - I still would prefer an 8″ 1024×600 screen. Not too expensive, but not as tiny as what the EeePC has to offer.

Other notable features include a touch screen display (this one is a convertible tablet design - I don’t really care for those to be honest; they mostly drive up weight), 40GB hard drive and Windows XP (oops). An AMD Geode L800 (which contrary to what the name might imply, runs at 500MHz) with 512MB RAM might not feel all that snappy running Windows - I think they should either have gone with more memory, or with a faster CPU (the Intel Celeron in the EeePC should give you much better performance), or stuck with Linux like many of their competitors. Or even better, all three.

And there’s another one. The Compaq 2133 pops up at all the usual sites - but interestingly not at HP’s own website. Oops.

So right now we are mainly dealing in rumors. The specs are promising (larger, higher resolution screen with 8.9″ and 1366×766 resolution - my main complaint about the EeePC is screen size), but the main question is price… the rumored $600 are way too high to attract people away from the EeePC, I think.

But I’m happy about all the competition in this segment. This year could be very interesting.

Lots of Cloudbook reviews are popping up all over the net. Here are the things that stick out for me:

• After all the complaining about the EeePC resolution, Cloudbook went with exactly the same 7″ screen with the same resolution. How unfortunate.
• gOS seems pretty but less mature. Several supports of hangs, of apps not working, generally of more user confusion. The EeePC user interface is very well done in comparison. The Firefox theme is better optimized for the small screen, etc.
• Generally the EeePC feels faster. Faster boot, faster application load times.
• Even though, some reviews compare the 1.2GHz Via C7 with the “900 MHz” Intel Celeron (guys, do your homework, it’s running at 630MHz) and assume that the C7 is faster. The reality is that it has a faster clock speed but is running about half as fast as the Celeron at the same clock - which means the two chips at question here are roughly the same, the one in the Asus maybe a tad faster - which the comparisons seem to show as well. And of course it’s trivial to push the EeePC to 900MHz (at the cost of some battery life). I guess I should mention here that I work for Intel, so get confirmation for this data elsewhere.
• Also, some reviews point out the 30GB hard disk in the Cloudbook as “better specs”; my only response to that is “excuse me”? I’ll take 4GB SSD over a 30GB hard disk in an ultra mobile notebook any time.
• Also, it seems that the chipset of the Cloudbook limits it to 1GB of RAM - I run my EeePC with 2GB and can tell you that alone makes a big difference.
• But the thing that really bugs me the most - the weird tiny touch pad in a very odd location on the Cloudbook (above the keyboard on the right). Not promising.

Overall I think that Asus doesn’t have to be too worried about this newest competitor. Especially as they are getting ready with their second revision that is generally expected to deal with the biggest complaint: screen size. The 8.9″ EeePC Gen2. Can’t wait to get one.

This is something that I wanted to work on for a while but I always got distrated. It’s still not perfect, but it works well enough that I want to share it here.

From what I understand this should work with all of the Linux based Motorola smartphones and actually might work with many other smartphones that can be programmed to work as USB modems. As written here it has been tested on a Motorola A1200 / Ming using the T-Mobile US EDGE network; pointers on how to use this on other carriers are at the end of this post.

Let’s start. You need an EeePC and a Motorola phone like the Ming and a USB cable. Make sure that the USB mode of the phone is set to USB Modem.

You also need to have wvdial and pppd installed on your EeePC. I might add a section on how to do this under the default Xandros version of Linux if people are interested; on EeeDora it’s just a yum install wvdial away.

On the EeePC you need to edit two files (I’ll simply assume that you know how to do this - there are many ways, for example sudo vi FILENAME)

/etc/ppp/options has only one line:
:192.168.0.254
which defines the default address to be used for the peer that the pppd will connect to.

/etv/wvdial/wvdial.conf takes a little more work:
[Dialer Defaults]
Baud = 460800
Modem Type = USB Modem
[Dialer MingUSB]
Modem = /dev/ttyACM0
Init1 = ATZ
Init2 = AT&FE0V1&C1S0=0
Init3 = AT+CGDCONT=1,\34IP\34,\34wap.voicestream.com\34,,0,0
Carrier Check = no
Stupid Mode = no
Phone = *99***1#
Username = tmobile
Password = tmobile
The Init3 string is one thing you need to change for different network providers (see below), the phone number seems to work for almost all carriers and and username and password appear to be arbitrary - Google is your friend if you run into a provider that actually checks them.

Now call wvdial USBMing (depending on the permissions in the distribution that you use, you may have to do this as root) and you should see something like this:
[hohndel@Eee linux-2.6]$ sudo wvdial MingUSB
–> WvDial: Internet dialer version 1.60
–> Initializing modem.
–> Sending: ATZ
ATZ
OK
–> Sending: AT&FE0V1&C1S0=0
AT&FE0V1&C1S0=0
OK
–> Sending: AT+CGDCONT=1,”IP”,”wap.voicestream.com”,,0,0
OK
–> Modem initialized.
–> Sending: ATDT*99***1#
–> Waiting for carrier.
CONNECT
–> Carrier detected. Waiting for prompt.
~[7f]}#@!}!}!} }< }!}$}%\}"}&} } } } }#}$@#}%}&h}#} } }'}"}(}"}#}<~
--> PPP negotiation detected.
–> Starting pppd at Sat Feb 16 21:20:06 2008
–> Pid of pppd: 2825
–> Using interface ppp0
–> pppd: �-8[08]h-8[08]
–> pppd: �-8[08]h-8[08]
–> pppd: �-8[08]h-8[08]
–> pppd: �-8[08]h-8[08]
–> local IP address 10.173.197.5
–> pppd: �-8[08]h-8[08]
–> remote IP address 192.168.0.254
–> pppd: �-8[08]h-8[08]
–> primary DNS address 66.94.9.120
–> pppd: �-8[08]h-8[08]
–> secondary DNS address 66.94.25.120
–> pppd: �-8[08]h-8[08]

The responses from the phone that look like garbage will likely differ when you try it; I included them here so you aren’t surprised when you see them.

There’s one problem remaining that I haven’t figured out, yet. After I kill wvdial in oder to interrupt the connection the phone doesn’t appear to interrupt the GPRS or EDGE connection to the network. A new attempt to connect will fail with a BUSY error. Unplugging and re-plugging the USB cable gets everything back to normal, but there has to be a better way to do this… please comment below if you have solution!

Finally, in order to use this with a different network provider, you need to change the the Init3 string. The format is AT+CGDCONT=1,\34IP\34,\34APN\34,,0,0; your provider will give you the Access Point Name (APN) to use and in most cases all you need to do is to insert that in the second set of double quotes (which are encoded as \34 to avoid cut and paste problems). Much more on this topic can be found in this article on GPRS, you may also want to look at a detailed explanation of the GPRS specific AT commands and finally at this PDF that has a lengthy list of possible settings (but interestingly enough doesn’t include T-Mobile US).