Sunday, August 07, 2011

Call to action for Verizon and Motorola

After about a month using the Droid 3 from Motorola / Verizon and comparing it to my prior Droid 1 (which was a Google Experience Device or GED), I’ve got some friendly advice for both Motorola and Verizon. First, here’s a look at the Verizon pre-installed bloatware:

Notice that most of it cannot be uninstalled, and some of it is actually running all the time.

OK, on to the advice:



Don’t install bloatware, junkware, trialware, etc. on our phones. For example this ridiculous “City ID” that pops up when you try to make a call asking if you want to pay for their silly service. It popped up on me while making a call in bright light out of doors. I think I answered it with the correct button for “Go Away”, but I could not really read it since phone displays are generally hard to read outside in bright light conditions. I just wanted to make a call. I don’t want to be interrupted by stupid trial ware. Honestly, if we wanted these applications we know how to use the Android Market or the Amazon AppStore for Android to go get them. Since you have chosen to make them un-removable, we can’t get the junk off of our phones. And, if there is an update to one of them in the market, the space for the pre-installed one is still used up since it cannot be overwritten. Here’s a couple of hints:

  • If you really think these applications are so great, put them in the Android Market and sell them or offer them to people who don’t have Verizon phones. If people really wanted them, you’d make money on them.
  • If you insist on installing them, they need to be un-installable. Can you imagine what would happen if say Dell or HP or Lenovo installed some crap that we could not remove?
  • Don’t try to trick people into using VZ Navigator at $9.99 a month when the free Google service meets almost everyone’s needs. Yes, you have a few features they don’t. But, for most people, nothing worth paying for.
  • People know they can create their own ringtones on their computer or using something like Ring Droid from the market. Don’t try to confuse them into buying them from you.

Do you think I want these applications cluttering my app drawer, using my memory, and sometimes running and using my limited RAM?

Not even a little bit.



(This first part about Android skins could apply to HTC with their Sense and Samsung with their TouchWiz as well). So, you don’t believe your hardware is enough of a differentiator. You believe you need to differentiate yourselves with a software “Skin” (MotoBlur or just Blur). Well, here’s some news on that: For technical phone users it just pisses us off and for non-technical folks they don’t even notice the difference. Like with some of these Verizon apps above – if you think Blur is a great differentiator, offer it in the Android Market for other devices and make a few bucks on it. If it is really that wonderful, some people running Nexus devices will buy it. Maybe a couple running HTC and Samsung devices might too. There would be your differentiator right there. The only problem? Oh, yeah – we hate Blur.

Looking at Blur, what does it bring to the table on the Droid 3?

  • Lots of bugs. Here’s just a couple of samples:
    • Won’t show pictures of contacts when dialing using Google Voice
    • AirPlane mode causes random reboots and if you have WiFi and BlueTooth off you can’t even exit AirPlane mode without rebooting
    • Camera app that doesn’t allow the user to set white balance for outdoors, fluorescent lights, incandescent lights, etc. and causes most indoor pictures to have a terrible blue tint.
    • SMS timestamps, even on restore from backup, are the time received and not the time sent
    • Home screen redraw lags (caused by Blur high memory usage and device low RAM)
    • Undocking from the media dock causes automatic brightness to be turned off
  • Nice widgets for favorite contacts and calendar
  • Slow performance

I don’t think anyone minds Motorola spending time developing Blur. However, you should separate this from the hardware device and either:

  • Pre-install it, but allow removal
  • Don’t install it at all, but put it in the Android Market. Possibly free for Motorola devices and with a charge for other devices?

Oh, and you should take the few widgets that people do like and make them into real widgets that can be used by all launchers. Sell them in the market if you want.

Both Verizon and Motorola

It is pretty clear with the popularity of replacement ROMs for other devices that users don’t like being tied to this pre-installed junk. We’d really like the ability to run AOSP (Android Open Source Project) deliverables plus the Google apps and Android Market. This should be easy for you to do. If you simply give us an option to choose “Blur + Verizon Bloat” (you can call it something like “Motorola and Verizon enhanced experience” as I know your marketers would want to promote it) or “AOSP” we’d mostly be happy (obviously there is a smaller niche who would still want custom ROMs but less folks would need to try unsupported / unsupportable methods if you offered a working AOSP we could use). Offering us AOSP + Market would allow those of us who want this better experience to have it without warranty killers like overclocking, etc. that come with some of the other ROMs.

Don’t just think about it – make this happen, and do it now.

Saturday, June 18, 2011

New Windows Home Server Installed

Over the last couple of weekends I installed a replacement Windows Home Server. I was moving from Windows Home Server v1 (based on Server 2003) running on an old Dell Dimension 8400 (a Pentium 4 660 running at 3.6 GHz) purchased in February of 2005 to Windows Home Server 2011 running on a newly purchased HP Pavilion 1080-CTO with a core i5-2500s quad core processor running at 2.7 GHz. Yes, the Dell was 6 and a half years old. It had been modified (by me) to hold 4 drives and was pulling between 228 and 288 Watts as measured by my UPS. It was getting to the point where that old space heater was going to die of old age and it needed to be replaced.

Since the new Home Server had recently shipped and was now available through NewEgg, I thought it would be a good time to upgrade from the also dated Windows Server 2003 based Home Server v1. I knew about Microsoft’s oft-decried decision to drop Drive Extender, but thought I could make do with hardware based SATA raid and just go with a pair of 2 TB drives in a mirror set (RAID 1). The old server had two 1 TB drives and 2 500 GB drives and it still had plenty of free space so this should have worked out fine, right? So what did I learn in this process?

First what I learned about HP’s consumer Pavilion line

HP makes it a bit difficult to figure out what you are buying and what is in it. For example, the machine I bought was sold as a “HP Pavilion p7qe series”. It said this on the page I ordered from, the receipt that they emailed me and the packing list that it came with. However, the support site will not admit that this machine exists. I figured I could wait a few days – as they said this was a new product and maybe the support site had not caught up. What I was wanting was to download the drivers required since I was not going to be running the supplied Windows 7 Home Premium. By the time the computer arrived, the support site still disavowed all knowledge of the p7qe’s existence. However, I managed to figure out that the support site wanted me to look up a “Pavilion 1080-CTO” instead. I did this and found only a modicum of drivers. Not even all the ones you would need to load a plain vanilla Windows OS on the machine. I downloaded the 7 or so that were there. It was obvious that HP doesn’t plan to even update these as they were all called “original driver” and appeared designed for you to use to revert to older drivers that came with the machine after some disaster in updating occurred. I powered on the machine, went through setup and proceeded to make sure all the supplied hardware worked. I also copied the c:\swsetup folder from the machine to a USB flash drive in case it had some of the drivers needed.

Everything worked, but it turns out that HP consumer level machines don’t bother to continue the very nice job their corporate focused line does on providing extra screws and mounting hardware. I had a second drive to mount, but could not mount it due to:

  • No power connector. The power supply in this box has 5 molex connectors, 2 SATA power connectors, and 1 floppy drive style connector. Of course the two SATA power connectors were in use by the existing hard drive and the DVD drive. Shipping a machine like this is silly as new drives do not have molex connectors. It is an option on the HP site to add a second optical drive (although why you would do so is a mystery). If you added it, they would have needed to use an adapter to connect the power. Fortunately a co-worker had such an adapter in his stash, so I used that.
  • No supplied mounting screws. I’m spoiled here by HPs fine corporate machines where the extra screws needed to add storage are nicely arrayed in the case by default. Not so on the consumer line. I scrounged several different types of screws from the old Dell and got the drive mounted.
  • No SATA cable. In the old IDE days you used one cable for the drive and most everyone shipped those cables with two connectors on it. In the new SATA days, I don’t think any of the vendors ship you a second cable. Had to get one from the old Dell.

One I got the drive mounted, I went into the “BIOS” (actually UEFI on this machine) and attempted to setup the RAID 1 mirror set. Alas, I could not do so. I’ve done this before on several different work machines so I didn’t think it was just the computer being smarter than me. I looked more closely at that support site now that I knew I could find it by the 1080-CTO moniker. I drilled down on the mainboard. I don’t know why they make their own mainboards instead of using an Intel or ASUS board, but apparently they do. This one was a “caramel” board. It has an Intel H61 chipset. No more information than that. So, I went to Intel’s site and much to my chagrin found that the H61 does not support SATA based RAID. Since I had been using that at work for about 5 years I didn’t think there would be any new chipsets that couldn’t do it but apparently there are. So, the SATA RAID was out. I decided automatic folder duplication would probably work for this and moved on. It wouldn’t protect the OS from a drive failure, but should work for the data.

I went ahead and installed Home Server. During the install it decided that I didn’t have a network connection and wanted me to load drivers. I fed it the USB key I had made and it was happy. Later I had a whole bunch of devices in Device Manager that needed drivers. These ran the gamut from video to sound. I tried the drivers I had downloaded from HP. Unfortunately, even though Windows 7 and Windows Server 2008 R2 (upon which Home Server 2011 is based) run the same drivers, HP had locked the installers so that they would only extract on Windows 7. Imagine going out of your way – spending money even – to prevent people from running free drivers on an operating system on which they work. Would you do it? Does it even make sense? I can’t see any way in which it makes sense for them to do that. It is almost like these crazy cell carriers who try to prevent you from using the features your phone software originally came with. Hell, if the machine wasn’t working and I wanted to extract the drivers on an old Vista machine so that I could put them on a memory key why should they stop me? Ridiculous. So it was off to Intel, RealTek, etc. to collect drivers. HP – fix this. Make drivers available so that a clean install of Windows has all the drivers it needs and don’t lock them to just a single flavor of an OS. You know as well as I do that Server 2008 R2 and Windows 7 use the same driver binaries.

What I learned about Home Server 2011

During the installation I mentioned that it wanted me to load network drivers. Fine, I did that. Then it would only allow a reboot. What? A reboot to load a network driver? I’m sorry – this install is based on Windows PE and you can dynamically load drivers all day in it. Even Windows Server 2003 on that old Dell which required a driver from floppy disk for the ICH6 based disk controller would load that storage driver and go on without a reboot. Not sure what someone at Microsoft was thinking there, but it smacks of just being lazy. Once I got the server up and running I realized that even though I knew Drive Extender was gone, that it also meant that automatic folder duplication – that amazing feature by which each data file is automatically written to at least 2 spindles – was also gone. I decided that I would go with some robocopy scripts to sync the data to the second drive for recovery purposes and resolutely moved on. At this point I was getting a bit frustrated with the new software since it was so many steps backwards, but was pretty committed to getting off of the ancient Server 2003 platform to something that at least supported SMBv2.

Next, I tried to access the Home Server from work. This was something I had done periodically before and it always worked well. I had my machines at home go into sleep mode, but I had written a Wake On LAN add-in for Home Server 2003 and used that to wake them if I needed something that was only available on my home machines and not just on the server. However, this new version of Home Server no longer lets you access the server “Console” from a remote session. (They call it the “Dashboard” now). But you cannot access the Dashboard remotely. After some digging, it looks like I need to write a “Gadget” for Wake On LAN that can be hosted in the Home Server web page. I’d have done that already but the SDK is a crock of crap. The documentation on MSDN isn’t in sync with the downloaded SDK at all – it tells you to install files that don’t exist (they are there, but with completely different names), tells you to do things in the wrong order, and doesn’t really provide a step by step “hello world” example that goes from start to deployed. Oh well, I will figure it out over time. For now, no more Wake On LAN for me.

Next I setup the backup. It was a simple matter to setup my robocopy scripts as scheduled tasks to do the work that automatic folder duplication should have been doing. But, I needed a real backup since there was no hardware based RAID. The new version has a backup – and it insanely requests that you run it twice a day. In fact, their documentation suggests that you tailor it to the needs of “your organization”. Did they forget that this is Windows Home Server and that your organization is your family? Anyway, I originally went with their recommendation for twice daily backups. Then I realized that the backups fail all the time, randomly. This appears to be a problem others share. The backup program was apparently some sort of last minute bolt on that doesn’t work well. In fact, with only 500 GB of data, the backup often takes 7 hours to complete. Imagine if I had maxed it out at 2 TB. Let’s see that would be 28 hours to do a backup. That’s when it doesn’t fail. I found that it fails almost all the time if it runs during the same window when client PC backups happen. Unfortunately the log will just tell you that at least one file wasn’t backed up successfully. It won’t list the ones with problems. So I cut the backups to once a day and moved them to 8 AM. Sometimes they finish at 3:00 PM. Other times they finish in 23 minutes. Other times they fail. Sometimes they run until 3:00 PM and still fail. This part of the product is really bad and needs a lot of help.

The new “Dashboard” is a mess. For example, if you are on the page that shows computers and backups while a backup of either a client or the server is happening the dashboard screen will flicker and redraw (changing your focus each time) so often that it becomes very difficult to right-click a machine and choose something from the context menu. I know it is written in .Net, but my god it is slow – even on a quad processor machine with 8 GB RAM. Apparently the developers don’t use the “InvalidateRect” function when they need to update their percentage complete on the backup and just redraw that small rectangle. Instead, they redraw the entire frame (which is silly and wasteful).

So, to sum up:

  • No drive extender
  • No folder duplication
  • Backup program is broken, and can only back up 2 TB max
  • No access to the “Console” (Dashboard) from the remote access page

All in all it looks to me like they are trying to kill off the product. Too bad to as it used to be really nice. If it wasn’t for the automatic backups of all the client computers, I would just go with a NAS setup like a Drobo or something.

Monday, February 07, 2011

Gawker sites latest to destroy themselves in search for elusive “new hotness”

I’m not sure what they were going for, but the Gawker sites (such as Gizmodo, LifeHacker, Jezebel, io9, etc.) have updated to a new format that is just unbearable. They used to have a fairly simple layout where you could just scroll down the page and see all of the articles / posts in a synopsis view and easily decide which of those were interesting to you, middle click the ones that were and they would open in a new tab for your reading pleasure. As all folks have different interests, I would typically find 1 in 6 interesting on say LifeHacker, maybe 1 in 8 on Gizmodo, etc.). Now, with their new layout, they have only one story per page. You have to go through a rigmarole of arrow keys to get to the next story – which probably isn’t even one you are interested in. They’ve added a “pane” on the right side that supposedly shows the latest stories, but lacks the size and impact of the previous blog style (pictures are tiny and synopsis too short). That pane on the right was taking over a minute to load earlier this morning, but it seems like maybe they have it fixed or at least implemented some sort of temporary workaround to get it to load more quickly.

Like the Digg users abandoned Digg in droves with the last redesign (I’ve rarely been back to that site which I used to use daily), it may be Gawker’s time to lose a lot of users. For now, they all seem to be supporting a “classic” view which, while not as good as their original view, is less of a pain than the new view. Classic doesn’t have the large size images and full synopsis that was available on the original sites.

I had tried to post my thoughts on the design as a comment on the Gizmodo article about the changes, but it looks like they removed Facebook Connect yet again (which you don’t find out until you have spent some time creating a thoughtful comment – only to find you can’t actually post it.

Here’s hoping that Engadget won’t follow the trend of making their properties less valuable to users; for now I can still visit that site and get a blog style view that works.