Skip to main content

@davewiner You're in Seattle next wk- I'd like to buy you a few beers or other drinks in thanks for all the snacks

@davewiner - Portland meetup sounds great but need a few hrs to get down there. Want to buy you a few in thanks for all the snacks & tools

RIP Mr. Garibaldi

1 min read

From Sleeping in Light: Our last look at Mr. Garibaldi before he boards the transport for the final time.  

It's sad to see so many Babylon 5 actors pass away so soon. Richard Biggs. Andreas Katsulas. Michael O'Hare. Jeff Conaway... and now Jerry Doyle. 

RIP Mr. Garibaldi.

Are e-books a scam?

1 min read


I was planning to re-read an e-book I purchased from Amazon back in 2014.  The book was missing from my Kindle and also missing from my online library. I searched for the book in the Amazon marketplace and they apparently don't carry it anymore.

I contacted Amazon and am looking forward to their explanation.

Have you ever encountered this issue with e-books?

Addendum:  Amazon restored the e-book for me.  There was no explanation about why it disappeared.  

I think I'll be going back to reading "real" books that can't disappear without an explanation. 




UX is not UI

3 min read

Recycling more bits from an old blog...

Many people seem to confuse UX (User Experience) with UI (User Interface).  These are very different things and confusing the two are likely to result in unhappy users. 

Many people who “design UX” seem to spend the bulk of their time focusing on the screen layout, controls and behavior of various elements on the screen. Many weeks can be spent conducting user studies, analyzing feedback and reviewing various fonts and color schemes in an attempt to provide the most aesthetically pleasing user experience.  These are important things to focus on but something obvious usually goes missing – what about the data?  Where is the data coming from?  How must the data be transformed or manipulated prior to displaying it to the user?  These critical questions are frequently an afterthought as design engineers focus instead upon the look and feel of a given screen.  Focusing on the look, layout and behavior of individual controls on the screen is UI, not UX.  UX should require that designers consider potential latencies due to data retrieval, security and other background processes that impact the overall performance of the user experience.  A beautifully designed screen that performs poorly always results in a poor user experience. 

So how should one design a great UX?  I’m not a UX expert but I’d suggest you start by understanding the data that the screen is to display or work with.  List all of the fields or data elements that the screen must display or accept from the user.  Try to answer the following questions:

  • Where is the data coming from?
  • How long will it take to retrieve the data to display to the user?
  • What relationships (if any) are inherent in the data to be displayed or processed?
  • Is the data in a format that is presentable to the user or must it be transformed?
  • For data entry try to invoke the validation rules as close as possible the actual data entry experience
  • Is it possible to retrieve the data proactively via a background process?
  • How can we “distract” the user while their data is being retrieved/transformed/validated or otherwise processed?

Avoid blocking, especially if the data must be retrieved or processed by a slow/remote network location or an underperforming back end system.  Users have become accustomed to asynchronous experiences – blocking users with an hourglass, spinner or some other form of “dancing monkey” is insufficient. 

Design engineers should work closely with architects or developers to become intimately familiar with the data and its associated service level expectations (SLEs).  Understanding data behaviors can help design engineers avoid designing a Ferrari that performs like a Yugo.


The Semantic Web is the Killer App for the Internet of Things

2 min read


Recycling some old bits...

I recently remarked how the future of the Internet of Things (IOT) was looking a bit like the original vision of the Semantic Web. This was based upon an old article on O'Reilly Radar.

The article painted a vision in which multiple devices are able to interact with each other due to the adoption of open standards. Open standards could one day enable an Internet of Things (IOT) in the same way the web itself grew and flourished due to the broad adoption of open standards.

This vision of devices interacting with one another reminded me of a fifteen year old Scientific American article by Tim Berners-Lee and Jim Hendler on the Semantic Web. That Scientific American article shared a scenario that was hinted at in Mike's O'Reilly article:

  • The entertainment system was belting out the Beatles' "We Can Work It Out" when the phone rang. When Pete answered, his phone turned the sound down by sending a message to all the other local devices that had a volume control.

Open standards are important to enable multiple devices to communicate with one another, but simply enabling communications may not be enough. As George Bernard Shaw so eloquently put it:

"The single biggest problem in communication is the illusion that it has taken place."

Without a common approach to semantic representation, enabling devices to communicate with one another will be a daunting task. As the Scientific American article indicates:

  • Semantic descriptions of device capabilities and functionality will let us achieve such automation with minimal human intervention. A trivial example occurs when Pete answers his phone and the stereo sound is turned down. Instead of having to program each specific appliance, he could program such a function once and for all to cover every local device that advertises having a volume control -- the TV, the DVD player and even the media players on the laptop that he brought home from work this one evening.

As the Internet of Things continues to evolve I hope that the massive semantic and security gaps can be properly addressed. The last thing anyone needs is an unpatchable device that won't communicate with anything, making it an overly-expensive doorstop waiting to be hacked.




Using webhooks in Logic Apps

4 min read

Update: This post was originally written before Logic Apps Designer fully supported webhooks.  Logic Apps Deisgner now fully supports webhooks.

New to webhooks? There is a short intro to webhooks in this post on DZone.  It does a decent job explaining the concept but more interestingly discusses how webhooks are composable.  Wonderful idea.

It's election season here in the US and that means polling.  Polling is fine for voting but is usually not a great idea for services.

If I've got a long-running process (e.g. waiting for a user to login and approve a document) I shouldn't have to poll to see if that process has completed yet. 

Polling is like going on a long drive with a kid in the back seat constantly asking "are we there yet? are we there yet?". 

Adopting an event-driven approach is much more natural and loosely coupled - the long-running process can simply tell us when it's complete instead of periodically polling it.  Being event-driven with RESTful services can be accomplished in many ways but one of the easiest is with webhooks.  A webhook is a user-defined HTTP callback.  I register my callback with the long-running process and wait for it to be invoked by a POST.  No polling needed. 

Webhooks are now supported by Logic Apps.  You can think of an Azure Logic App as a workflow that you build and run in the cloud. With webhooks your workflow can wait for an event from a long-running process. 

Here is a simple Logic App that consists of two steps.  The first step POSTs a simple message to a RequestBin URL. Request Bin is a great, free service you can use to test and inspect HTTP messaging.  The second step is another HTTP action but it has a strange icon on it - it also doesn't render properly in the designer.  This is because webhooks aren’t yet supported in the Logic Apps Designer (this is coming soon).  Webhooks are supported in the code view so let’s look at some code. 

Here is (most of) the code view of our simple Logic App (Logic Apps use JSON under the covers). There are two steps in this workflow.  The first step is named "Http" (default name - I could have changed it but I didn't).  The second step is named "Http_2".  You'll see each step contains an optional set of conditions, inputs and a type identifier.  

Note the second step of the workflow ("Http_2") surrounded by the red box.   Looking at this code we can see some interesting things...

If you look at the inputs you'll see it contains a subscribe construct.  The @listCallbackUrl() enables us to retrieve the URL for firing the webhook (where the long running process can "publish" its event to).  The type has been changed from Http to HttpWebhook.  The URL being used here came from RequestBin.  These changes had to be made manually in code view because the designer doesn't yet support webhooks.  This "Http_2" step originally looked like the "Http" step right above it.

What happens when we run the workflow?  It stops and waits for the webhook we subscribed to:

If we go over to RequestBin we can copy the webhook URL (in red below) and use it to POST back using a tool like Postman or Fiddler (click to enlarge):

Now that the webhook has fired (was POSTed back) the workflow will wake up and continue running:

Learn more about the JSON and commands used by Logic Apps here.

Note: Old-timers like me might be surprised to see "Workflow Definition Language" making a comeback.  Despite the similar acronym this is not the same WIDL that was used by the now defunct webMethods product.  



Setting Up and Running Your Own River 5 Server on Azure (For Poets)

13 min read

Note: Click on thumbnails for a larger/readable image.

Dave Winer is brilliant guy who made RSS a standard,  "bootstrapped the blogging revolution" and invented a ton of incredibly elegant and useful tools.  If you're not reading his blog you're seriously missing out.

One of the many interesting ideas he developed was the "river of news". A "river of news" is different from standard news readers and provides a much better user experience (don't take my word for it, go read about it yourself).  Dave recently shipped River 5, the latest implementation of the "river of news". 

There are some very good instructions for installing River 5 on a client (Mac, Windows) but I haven't found any clear instructions for creating a River 5 server, hence this post.  (I'm also a big fan of Dave's "For Poets" series - this is my feeble attempt at creating such a post.)

OK - let's get started!

You'll Need A Server

Running a server obviously requires you to have a server, right?  Not necessarily - this is the age of the "cloud" and you can run your server "in the cloud" without having to buy it and maintain it.  The "cloud" is just a trendy term for using someone else's computers for your own stuff.  In this case someone else's computers will be Microsoft's Azure service and your own stuff will be River 5. 

Wait - Microsoft is evil and proprietary isn't it?.

This is not the same old Microsoft. Azure supports several operating systems (not just Windows) and open source tools. Many Microsoft products and frameworks are now open source and available on Git.

Step One: Get an Azure subscription

This is the easiest step.  Best of all it will enable you to explore the range of services and capabilities available in Azure once you're more comfortable with it.  Signing up is easy.  You'll even get $200 to spend so it won't cost you anything to set up your River5 server.  Now go get signed up and come back here when you're ready.

Step Two: Pick a server

Remember, the cloud is nothing more than someone else's set of computers that you can use for your own stuff. Here's where we pick what type of computer we'll be using for our River 5 server.

Now that you have an Azure subscription, log into the Azure Management Portal

When you first sign into Azure it may look a little intimidating because of all the stuff you can do.  Don't worry about all that stuff at the moment, you can explore later.  For now, click the + New and click Compute.  You'll see a wide range or servers you can start up and use.  As you can see it's not just Windows Servers.  These are pre-configured special purpose servers designed for databases, SharePoint and more. 

Click for a larger view

Your River 5 server will use the latest and greatest version of Windows Server: Windows Server 2016.  Don't pick the one you see in the menu though, that's a much beefier server than you'll need for River 5. We're going to pick something simpler to use for our River 5 server.  Click the See all blue text in the upper right corner.

Now click the Search box, type Windows Server 2016 and hit ENTER.  Now pick Windows Server 2016 Technical Preview 4. 

Click for a larger view

Another window will appear asking you about deployment model options.  Accept the default setting and click Create.

Click for a larger view

Step Three: Configure, start up and login to your server

Enter your server name and define a userid and password (make a note of these - you'll need them to login soon).  Select your Azure subscriptoin (created in the first step) and define a Resource Group. An Azure Resource Group is like a bucket containing all the information needed to manage the Azure service you're using. In our case the Resource Group will contain information about your server and it's networking capabilites.  Lastly, pick a location where your server will be hosted.  You should pick a location nearest you.

Once you've filled in this information click the OK button.

Click for a larger view

Now we need to select the size of our server.  There are many server sizes available in Azure - from simple and cheap to highly complex and expensive.  You'll initially be shown some recommended server sizes.  Ignore the recommendations and click the View all blue text in the upper right corner.

Click for a larger view

Server sizes are identified by a letter/number combination but you can also see each the specifications associated with each server size.  For this exercise we're going to pick the simplest server available: the A0.  Scroll down the list until you locate the A0 then click on it.  (Note: You'll see an estimated monthly cost for the server. You received a free $200 credit when you signed up so going through this exercise is free for you.)

Now click the Select button.

Click for a larger view

Back in the main menu again, pick the Public IP option, then pick Configuration.  Here's where we give your server a friendly name so you won't have to type in an IP address to reach it.  In this example I used "River5".   The address you'll enter into your browser will be the name you enter plus the additional text that appears below it.  In this example the address for my server will be  (:1337 is the port used by River5 - more about that in our next step).  Make a note of your URL - you're going to use it later to access your River of news.

Now click Save at the top of the screen.


Now we're going to open up the ports that your server will use to communicate with a browser like Chrome or Internet Explorer.  Click Network Security Group then click Inbound Security Rules. Add and save two inbound rules: one for port 80 and one for port 1337 as shown below.  Remember to click the Save button after entering each port!


We'll use the defaults for the rest of the options.  Now click the OK button and you'll see a summary of your server settings (on the far right): 


Click the OK button.  The setup will close and you'll land back on the Azure Management page. See that blue box?  That's Azure deploying and starting up your new server for you.  This may take a minute or so.

Click for a larger view

When your server is ready the blue box will change and look something like this:

Click for a larger view

Clicking on this box shows you information about your server.  Click Connect from the menu at the top.

Click for a larger view

This will download the configuration to remotely log into to your server.  Save this file to your desktop and double-click on it.  You'll see the following warning message because you've never connected to this server before. Click Yes to continue to the login prompt for your new server. 

Click for a larger view

Login with the userid and password you defined in step three above.  After a delay your new server's desktop should appear.  Congratulations!  You now have your very own server! 

Step Four: Server Housekeeping

We're almost ready to install River5. We need to clean up a few things to make your new server easier to use.  Servers are very different from desktops machines - they are typically locked down to be secure.  We're going to tweak some of these security settings to make your server easier for you to use.

If the Server Manager wizard appeared after you logged in, close it by clicking the "X" in the upper right corner.

Click to enlarge

First we're going to tweak a security setting that will enable you use the default browser known as Edge (servers are normally tightly locked down for security).  Click the Windows logo in the lower left corner, type RUN and hit enter.  This will pop up a command box.  Type SECPOL.MSC into the command box and hit enter.  The Local Security Policy Manager window will appear.

Click the Local Policies folder on the left and double-click Security Options on the right pane.  Now scroll down until you see the User Account Control: Admin Approval Mode for the Built-In... option (as highlighted in the picture below).  Double-click on it.

Click to enlarge

Click Enable then click OK to save.  Click the X in the upper right of the Local Security Policy Manager to close it.   

We're eventually going to have to restart your server to recognize the new security policy but not just yet.  Remember how we set up the ports in a previous step?  We're going to have to set up those ports in Windows Firewall to ensure they're working properly. 

Hold down the Windows key and press X.  Click on Control Panel in the pop-up menu.  In Control Panel change the View By option to Small Icons. Now click Windows Firewall and the firewall configuration window will appear.  Click Change Notification Settings on the left side of the window.  Click to be notified when Firewall blocks an app, as shown below:

Click to embiggen

Now we need to open the ports we set up in Azure.  Click the Advanced Settings from the left menu and click Inbound Rules.  We're going to set up two rules for the two ports we opened up in Azure.  Click New Rule from the menu on the right.  Click Port.

Click to enlarge

Leave all the defaults and enter 1337 for the specific port.  This is the port used by River5.  Click Next, select Allow the Connection and cick Next.

Click to enlarge

Leave the defaults and click Next

Click to enlarge

Assign the connection a name and click OK to save it.  You should also set up another Inbound rule for Port 80

Click to enlarge

Your server housekeeping is now done!  We just need to reboot the server because of the security policy change we made earlier.  To do this click the Windows button (lower left corner of the screen), click Power then click Restart. If you get prompted for a reason click Unplanned and continue for the restart to occur.  You'll lose your connection to the server when it restarts (the window will close).

Install and Set Up River 5 on Your New Server

Double-click on the configuration file you previously downloaded.  Log back into your server.  If the Server Manager wizard appeas click the "X" in the upper right corner to close it.

First we're going to download and install node.js.  Node.js is a server-based Javascript framework used by River5.  Start up the Edge web browser (blue "e" on the bottom left of the window) and download the installer for Windows.  After the installer finishes you can ensure node is working properly by:

  1. Hold down the Windows key and press R to pop up the Run prompt.  Type cmd and hit ENTER.  The Windows command line should appear.   
  2. Type node -v and hit enter.  Information about the version of node should appear.
  3. Type npm -version and hit enter.  Information about the version of the Node Package Manager should appear.

We're finally ready to install River5!  Using the Edge browser, download the River5 zip file

Right-click on the file and select Extract All.  You'll be prompted for the location you'd like to extract the files. I chose C:\Tools\River5 but you're welcome to choose whatever location you want. 

Hold down the Windows key and press R to open the Run prompt.  Type CMD and hit ENTER to open the command prompt again.

Navigate the folder where you extracted the River5 files.  If you extracted the River5 files to C:\Tools\River5 use the following commands:

  • cd \ and hit ENTER to go to the root of the C: drive
  • cd \Tools\River5 and hit ENTER

Once you're in this folder type npm install and hit ENTER.  This will install River5 on your server for you.

After the install completes running type the following command and hit ENTER to start River5: node river5.js.  You'll see some information scroll by about various news sites.

Open the Edge browser (on your server) and navigate to http://localhost:1337.  It may take a while for some news feeds to appear so be patient.  Eventually you will start seeing news feeds show up when you refresh the screen.

Set River5 to Start Automatically

You can set up River5 to automatically start in case your sever ever gets rebooted. This can be helpful in case of unexpected outages or restarts. First, create a simple batch file to run River5 by copying these commands and saving them into a file named run.cmd (note that you'll have to change the location of the River5 files if it differs from C:\tools\river5):

   @echo off
   cd \
   cd C:\tools\river5
   set NodePackagesPath=C:\tools\river5\node_modules 
   set Path=C:\tools\river5;C:\tools\river5\node_modules\.bin;%PATH%
   set Path=%NodePackagesPath%;%PATH%
   set NODE_PATH=%NodePackagesPath%\node_modules;%NODE_PATH%
   node.exe river5.js

Hold down the Windows key and press R to open the Run prompt.  Type regedit and press ENTER to open the Registry Editor.  The Registry is a special configuration that Windows uses to store information about the various programs you have installed.

Click to enlarge

On the left side click HKEY_CURRENT_USER, then click SOFTWARE and scroll down to MICROSOFT.

Scroll down to Windows and double-click it to open the tree structure underneath it.

Look underneath the Windows folder, locate the CurrentVersion folder and double-click it to reveal the tree structure beneath it.

Scroll down to Run and single-click it.

In the right side you'll see a few folders. Right-click in an empty area and click New.  Click String Value.  A new item should appear on the right side. Name it River5 and hit ENTER.

Double-click on the River5 entry you just created.  Type the full location of run.cmd into the Value box (e.g. C:\tools\river5\run.cmd).

You're finished.  Close the Registry Editor by clicking the X in the upper right corner.

You can also close the window for your server. It's ready to use!

Next Steps

You can now access your River from anywhere using the URL you noted previously (mine is at

See Dave's post for directions on configuring your own specific news feeds and other options. Pretty much eveything is configurable.






Enabling deep linking on my blog

1 min read

I recently added AnchorJS to my blog.  This enables me to add deep linking to any paragraph.  AnchorJS is a simple Javascript library that modifies any format tag you like by embedding an anchor, enabling you to deep link to it. I've enabled it on the paragraph tag and may tweak it a bit.

I first saw this behavior on Dave Winer's wonderful Scripting News blog and always wanted to enable it here.

Now I just need to blog more.  :)