Thursday, July 20, 2017

Reading on the iPad Part 13: New version of WSJ iPad app wrecked the user experience.

First a tip of the hat to my lucky number thirteen. Born on the 13th, and mom now living at #13, thirteen been very very good to me. Alas, Thursday, 13 July 2017 was not a lucky day for my reading experience with the Wall Street Journal iPad app.

Like installment 12, this review is a candidate for my all negative all the time blog, Angry Poetnerd. On 13 July, for my daily read of the Wall Street Journal, I was greeted with a perky, self-congratulatory pop-up announcing the new app version. I hate the new user interface. This article details why.

As early as 2011, in Reading on the iPad Part 4: Extreme disappointment: The Wall Street Journal, I expressed disappointment. I found basic usability problems. Although there were a couple areas where I changed my mind, and where the app improved, I continue to feel that the WSJ iPad developers fail to understand the needs of daily readers. I feel they systematically make a bad app worse -- failing to fix basic problems, and introducing aspects that harm usability.

Review of initial problems:

  • Failure to adopt iPad standard "draw" and "pinch" gestures to enlarge and shrink image. Never fixed. The app still wastes precious screen space with a font resize button. Worse, you don't choose the size. You step through 3 hard-coded sizes.
  • Co-opting standard "pinch" gesture to close article and go back. Fixed. A Back button was added.
  • Broadsheet layout retained, but squished into iPad form factor resulting in tiny article introductions that usually end in the middle of a sentence. Accepted. My opinion here has changed from initial hatred to approval. The positioning and relative sizes of articles turn out to be useful typographical cues. Having seen the Washington Post app abandon this, I miss it.
  • No "Table of Contents View". Fixed for 5 years; broken this week. More details below.

Buggy operation around suspend and sleep:

In my early review, I encountered many problems around resuming the app after either running another app, or putting the iPad to sleep for a while while I did something else. I chose not to mention these problems initially because I expected them to be fixed in short order, and because I wanted to focus on the main flow through the app.

In retrospect documenting the bad behavior from the beginning might have provided useful documentation to other users, and might have gotten usefully back to the developers. Here it is, six years later, and the Wall Street Journal app still simply dies sometimes when I put my iPad to sleep.
It still sometimes completely forgets where I was when I go back to the app. It still sometimes resumes with chunks of the page simply missing. When re-opening the app on a new day, it still often fails to offer a pop-up to take you to the new issue.

No other iPad reading app that I have used has this much difficulty handling resume from suspend and sleep.

Evolving/Devolving Workflow:

In my second post of this series, I provided background and the evaluation criteria for apps that I would review. I suggested that two interfaces were necessary for a complete solution: one that let you "turn all the pages," and one that let you "navigate a table of contents view." In my initial review of the WSJ app, I lamented the lack of a table of contents view.

Shortly thereafter a table of contents view was added. In retrospect, I should have provided an update so as to give credit where credit was due, and to document the evolution of the interface.

The view worked like this: Tap on the three bars in the upper right hand corner to get a list of sections. When you tapped on a section, you got the front page of that section. You could tap on the three bars again to see a list of the articles in the section. It was not ideal for me flipping between a page display and then having to re-enter the table of contents.

Somewhere in 2014, if I'm properly interpreting my somewhat incomplete notes, I spoke with the Digital Editor and expressed how much I hated the table of contents view. I was told that improvements were coming, and indeed one day they came.

The interface changed, and I liked it. Instead of getting the first page of the section, you got the list of articles for the section. For me this was the ideal way to navigate.

  • I always knew all the sections that were available.
  • With a single finger tap, I could see the articles in the section.
  • It was easy to know how many sections and how many articles there were so I could plan my read-through.
  • The app was good at returning me to the table of contents with the article I just read highlit, so it was easy to keep my place as I stepped through the articles.

I didn't make note of the exact date and the exact workflow, because I figured that once the good workflow was established its value would be obvious and it would be kept. On 26 October 2016, an update came through that went back to the old "dump you in the first page of the section" interface. A new bug: sometimes the highlighting failed.

Damaged workflow detailed:

The update that came on 13 July, however, so totally changed the interface, that it provoked me to invest the time to produce this article. I am completely mystified how anyone could consider the new interface an acceptable way to navigate the content.

It used to be that you always had the 3 bars to take you to the list of sections. One of the sections was "Issues". From that page, you could select either the "Latest News" or a particular day's edition of the paper. When the notifier pop-up for a new issue failed, as it often did, it was easy to recover by going to "Issues" and selecting the new one.

Now when you start the app you land on the "Latest News" section by default. The three bars are gone, but now you have "Menu". I think "Menu" is more obvious than three bars, but what is under the menu is not at all obvious. How do you get to today's paper? You just have to know to tap on "Issues". From there you get a scrolling display of the different issues. What was once a mild inconvenience when the notifier bug bit, is now the normal workflow. Where once I got today's paper by default, I now tap and look tap and look. Annoying.

It used to be that, when viewing the table of contents view of a particular edition, you got additional sections like "Journal Report" and "Wall Street Journal Magazine." This made perfect sense. Just like how you received special sections in the print edition, there they were in the iPad edition. With the new interface, you don't get those. They're now grouped ONLY in the "Latest News" edition's table of contents. Since I never read the "Latest News" version, I missed a couple of these inserts until I figured out I needed to start looking there. What I got by default in a timely and graceful manner is now off someplace I used never to look, and have to remember to visit. More annoying.

Having to open the list of articles view again for every section was annoying. But the new interface solution is worse by far. There is no list of articles. You just have to know that you scroll down to see the articles in the section, and you scroll to the right to see the other sections. Are there indicators of any kind to let you know there are more articles to see? No. Well there might be. Sometimes I see a big bold "V" at the bottom of the section screen. But sometimes it's not there when there is more to see, and sometimes it's there when there is nothing more to see. Sometimes I scroll down and get a totally blank page. What used to be a simple, obvious display of what articles there were, how many there were, where you left off in your reading, is all gone, and replaced with an interface that has to be explained and which fails to provide useful, consistent indication of where to go next.

Having had a really good experience with the 2014/2015 edition of the table of contents view, I am beside myself in hatred of this new interface. I am seriously considering canceling my WSJ subscription and switching to reading the Financial Times via Zinio.

To anyone who is reading this article and who also reads the Wall Street Journal on the iPad, I'd very much like to hear if you agree or disagree with me here. Definitely call up the Journal and provide your feedback. When I went to the reviews on the App Store, it seemed like the negative reviews of the new interface outnumbered the positive ones by like 5 to one or more.

Continued Struggles:

I'm not sure if readers will find it easiest to get updates at the end of this article or with new postings. Please feel free to comment with what works better for you.

28 July 2017: The reason why broken recovery from sleep is so annoying:

Breakfast today (when I read the WSJ) has been interrupted by various things. I was in the middle of reading the article "Scaramucci Erupts Over Priebus, Leakers" in today's paper, and I stepped away from my iPad.

I fingerprinted back in, and landed not in the app, but in the folder containing the app. I guess the app crashed.

I restart it, and instead of going to the article I was in the middle of, I got dumped into the front page of "Latest News." To get back to where I was, I do this:
  1. Tap Menu
  2. Tap "Issues"
  3. Tap on the issue I was reading, Friday 28 July 2017.
  4. Tap on "Page One"
  5. Tap on the article I was reading, "Scaramucci Erupts Over Priebus, Leakers"
There. Wasn't that easy? No it wasn't! It epitomizes how the new version has mis-organized the content, and deleted useful interfaces into the content so that the broken resume from sleep forces me to tap tap tap just to get back to where I was in reading today's paper. This is the fundamental problem I have with the new version.

Contrast this with how I would recover in the previous interface I liked best but which got harmed with the update of October 2016:

Restarting the app brings me to the issue I was previously reading.
If I was on the front page, tap on the article. Done.


  1. Tap on the 3 bars to bring up the list of sections.
  2. Tap on the section I was reading.
  3. From the list of articles presented, tap on the one I wanted. Done.

With the October 2016 reversion to the "Force the reader onto the first page of the section and make them re-open the menu" interface, those 3 steps becomes 4 steps:

  1. Tap on the 3 bars to bring up the list of sections.
  2. Tap on the section I was reading.
  3. Tap on the 3 bars to bring up the list of articles.
  4. From the list of articles presented, tap on the one I wanted. Done.
Sometimes one step. Sometimes 3 or 4 steps. Now always 5 steps. This is bad workflow design and represents ignorance of user experience. (In addition to failure, across many years, to get "Restart after sleep" working.)

Thursday, July 13, 2017

How Comcast SHOULD have worked to fix signal leakage problems, but DIDN'T.

Everybody whose worked with a cable or telephone company has a story to tell about crappy customer service. Recently Comcast called saying there was trouble anticipated on my line and asking to fix it. What should have been a simple proactive service call has turned into a multi-call multi-day fiasco. It did not have to be this way.

I claim some blame for getting angry and not responding in the most constructive way. I offer suggestions in how Comcast could have communicated better through the multiple interactions that could have avoided my anger, and put things back on track.

The story/problem:

Upon my return from vacation I found multiple messages on my phone answering machine, from someone identifying themselves as calling from Comcast, saying that their monitoring system had identified a problem with my network and asking me to call 844-482-6747 between 8AM and 8PM to work on the problem.

An internet search on that number turned up several forum posts saying they believed the call was a scam. The CallerID on my phone reported either "Unavailable" or "Private Out of Area". Signing onto the Comcast web site and reviewing the calls log showed, "Cell Phone MA (508) 922-7953". Searching for that number again turned up much speculation that it was a scam.

I called my the Comcast service number that was on my bill, and asked about it. Surprise! It was legitimate. I was asked to schedule a service call, but that the work was strictly to be outside my house and I didn't even need to be home for it. When I asked why bother scheduling a call, I was told quite confidently by representative Anne that it was so I would know why a Comcast truck was outside of my house. The visit was scheduled for between 10 and 12 a couple days later.

At 9:30 the Tech shows up and wants to come in and check for loose wiring. I confess I did not respond constructively to this, and said angrily, "That's not what I was told!" My partner, a senior network engineer, offered to do the testing. I said a bunch of things in a hostile tone including, "Do what you have to do," "Perform your outside testing," and "We've had the line to the house replaced five years ago when squirrels bit through it. Check that." The Tech simply left and logged that he was not allowed in the house. That log information didn't reach me until the third follow-up call I made asking what the resolution was.

I called Comcast in a rage because I didn't want my service to go bad, but that I had not been communicated with properly. The first representative I spoke with started working through a script, and I hung up. The next representative I spoke with offered to log a feedback ticket, with an ID could use for check-ins, documenting how I was told the wrong thing about the nature of the service call, and that the tech left without saying what the resolution was.

Later that evening I got another call from Comcast just like the ones I got while on vacation. I called the 844 number, and said, "The Tech has already been here. But I do not know if the problem was or was not resolved." The caller apologized, and said that after reviewing the information I could be assured the problem was resolved.

I called my regular service number, referenced the ticket, and asked what the story was. I was told, "The real nature of the problem was a problem with your internal wiring, and we need to schedule a tech to come and look at it." This being the fourth different story, I got really angry and demanded to talk to a supervisor. I had to get a bit rude to get that demand satisfied.

To the supervisor, I inventoried all the wrong stories I had gotten thus far, and he said a new visit had been scheduled. Again I got angry, because I had not agreed to any new visit. A visit on Sunday was offerend (which would normally be quite nice) but I was not going to be home on Sunday, and I was angry that I seemed to be getting told, not asked about a service visit. We agreed to a different day later in the week.

This morning I got another call from Comcast, this time with valid Caller ID asking to set up a visit to work on a network problem. I angrily quoted the ticket ID and said, there is already a visit scheduled.

So the problems were:

  • Incorrect description of work needed and nature of service call.
  • Inability to come to understanding with visiting technician.
  • Delay in getting information on why technician left back to me.
  • Disconnect between service call setup and service call requests.
  • Incorrect statement that problem was resolved.
  • Continued disconnect between service call setup and service call requests.

Again, I admit to not having responded in the most constructive ways to these. The result was that I wasted my own time, and probably made it harder for Comcast to do the right thing.

If Comcast had told me up front, that a technician needed to inspect wiring in the house, or if I had been less hostile to that change of plan when the technician had come, the problem would have been resolved quickly and easily.

However, I would suggest that the internal fragile processes within Comcast caused several missed opportunities to make this all run more smoothly:

The solution:

Make the service process more robust by understanding the nature of failed communication. Add a little bit here and there to reduce mis-communication and to respond constructively to it when the inevitable mis-communication happens.

1. If you have a proactive trouble identification system, give clear indication of how to confirm it is not a scam. Adding, "or call the Comcast service telephone number on your bill." to the message would have given a quick and easy way to verify the call was legitimate. Don't hide your Caller ID!

2. When you are doing proactive trouble resolution, give the people working the issue the CORRECT information to communicate to the customer.

3. When you send your service technicians out to the field for proactive work, train them in responding to customers taken by surprise in mis-communication of the nature of the call.

4. Connect your proactive work order system to your regular service order system.

5. Remind your service personnel to AVOID MAKING STUFF UP when they don't know the answer. Make sure the correct answer is made available to them to communicate. Make sure they understand the customer's attempt to resolve conflicting information coming from multiple sources.

As things stand, the Comcast system for proactively correcting cable network leakage problems is fragile, and can too easily turn into a system for creating angry customers. Indeed I more than once said, "Maybe I should switch to FiOS." I'm lucky to have a choice. But I'd prefer for both me and Comcast to win here going forward.

Let's see if the problem does get properly diagnosed and fixed next week.


  • The technician showed up within the scheduled timeframe.
  • A quick connection to the house wiring from outside of the house determined that the RF leakage was acceptably low.
  • The technician marked the issue as resolved.
  • The supervisor who scheduled this visit promised I would get a call half an hour before the technician showed up.  None was received.
  • The supervisor promised he would call on the afternoon of the visit to check in.  No call has been received.
When the first technician visited, we ourselves found and corrected one imperfectly tightened down connection behind a switch plate.  We also realized we had one connection in the house wiring disconnected while we were away on vacation.  We theorized that the disconnection was what Comcast detected, but multiple calls, and multiple visits were required to get us to the point of recognizing that the current state of affairs is just fine.

There was no email in my Comcast inbox confirming either of the two scheduled visits.

One interpretation of events is, "If the first technician had done his job and checked the outside wiring as was originally agreed, everything would have just run smoothly."  However I'd say that my hostile reception for the first technician made communication too challenging for him. If he had said, "I checked the outside wiring as you were told, but I can show you how there is noise coming into Comcast lines from inside your house," I would have let him in.  Maybe we'd have found the loose connection as root cause and fixed it together.  But the first technician could not know that I was receptive to that message.

However, that situation would not have arisen if basic communication had been done right. Instead, communication of:
  • What the problem is.
  • How we will know the problem is solved.
  • The conditions on which inspection of inside wiring will be needed.
  • Consistent status reporting.
  • Promised follow-up contacts.
Were ALL failures.  0 out of 10 for communication, Comcast.


I finally did get the promised follow-up call from the Supervisor.  We reviewed the other communication failures. He told me that the problem arose on July 1, the day we disconnected the coax from our line filter. This tilts my belief in the direction of, "We caused this, and fixed it ourselves, but didn't have a functional conversation with the first technician to resolve the situation."

I'll work harder myself to inform, rather than act out displeasure when I'm confronted by technicians doing unexpected stuff. I had the opportunity to make this iffy situation better and I blew it.  I hope that Comcast also learns from this and does better with explaining the problem, and working through it.

Friday, June 30, 2017

Reading on the iPad Part 12: Drop Fast Company's app. They have.

I have another blog, Angry Poetnerd for posts where I have only bad things to say. By rights this posting should go to that blog.

With the May 2017 issue, Fast Company stopped doing any extra work to re-render for the iPad app. Instead the regular printed version was loaded verbatim into the app with no post processing, and no additional work on the app to deal with the new requirements.

  • Small, unreadable font: No ability to enlarge the font was added, so you were stuck with the large form factor magazine page in the small form factor iPad screen. This is a problem for near-sighted readers like myself.
  • Links needed but not supported: No links were included to go to or get back from articles that were continued elsewhere in the issue.
  • No additional navigation aids offered: To get to a continuation or back from it, you use trial and error either with the scroll bar and page flipping.

I Emailed Customer Care and the Letter to the Editor addresses raising my concerns. I never heard back from the Editor, but Customer Care was happy to refund the unused fraction of my subscription so that I could switch to Zinio.

I appreciate that creating an iPad-specific design and rendering is expensive. I'm personally skeptical of its value. But the way Fast Company transitioned out of this model was terrible. The change was un-announced and completely ruined the reading experience for me. I am only continuing as a subscriber because of the availablity of Fast Company via Zinio.  Zinio offers an excellent balance of cost and benefit for reading on the iPad. See also: Reading on the iPad Part 11: Zinio, My New Favorite

I continue to highly recommend that magazine publishers switch to Zinio. For example the IEEE. qmags still hasn't fixed any of the problems I flagged years ago. They have been surpassed by Zinio, but, alas, the IEEE has not yet noticed.

In related news:

The IEEE Computer Society has quit offering IEEE Computer via qmags. Sadly, they didn't adopt Zinio. There was a self congratulatory note sent out touting the additional formats available instead. For example, Kindle Format. Alas, the Kindle rendering was as bad as the Forbes rendering I reviewed back in 2011 (Reading on the iPad Part 6: Forbes was a disappointment.) I summarize the relevant bit:

The beautiful glossy magazine had been ground down into a cheap paperback book. The pages shrank down to 3x6 inches and the typography was reduced to a single font in only a couple point sizes. Colorful icons and graphical navigational aids were eliminated and everything re-formatted into block paragraphs. But as if to comply with a marketing directive to be able to say, "We offer color!" a very few select photographs were retained as "color plates" scattered throughout.

Forbes sensibly abandoned that format. The IEEE Computer Society has yet to learn what was obvious to some of us in 2011.

Pretty much the only rendering of IEEE Computer I find useful is the PDF format. With a few clicks I can get from the Email announcing the new issue through the iPad web browser into iBooks. If only Apple would copy Zinio's great interface for organizing magazine issues and bookmarks across items.

I feel that the IEEE Computer Society got caught up in offering a list of features, and lost contact with how to provide the best user experience. In my opinion, inadequate outreach was made to their own contacts in the User Experience realm.

Monday, June 26, 2017

PiDP-8 USB Hack

I have just finished building a PiDP-8/I from a fantastic kit offered by Oscar (Obsolescence Guaranteed) Vermeulen. This is a 2/3 scale reproduction of the front panel of the Digital Equipment Corporation PDP-8/I computer. Inside is a Raspberry Pi using SIMH (See also: SIMH project on GitHub.) to emulate a PDP-8. The PDP-8/I has a special place in my heart as the first computer I really got to know. You always remember your first...

One of the challenges is to bring the various and sundry connections through the bamboo case out to the world in a pleasing manner. I managed this:

Details on the connectors:

The power, audio, and HDMI extenders were findable with care.

Power was pretty straightforward.  I went with one from Renrii sold through Amazon, but I see it is no longer available, only a month later. Word of advice, look closely at the pictures, "Left Angle" and "Right Angle" are used inconsistently from vendor to vendor.

Audio extender to the box was straightforward.  I picked up this one from Amazon:  [Benly-38] Right Angle 4-pole 1/8" 3.5mm Stereo Plug / Male to Straight 4-pole 1/8" 3.5mm Stereo Jack / Female Aux Headphone Cable 0.55 Feet (17cm) Wvyr-4l4sf-15

But once outside the box, if you want to split the 4 pole 1/8" jack into separate RCA connectors, you must buy VERY carefully.  The assignments are not standardized, and some of the splitters have ground where you need signal.  My recommendation:  Buy new/old stock Zune a/v output cable. I got mine from a third party seller on Amazon:  Zune A/V Output Cable (Discontinued by Manufacturer). I confess I've not tested this yet, so I'm not 100% sure this is right.

For HDMI, I tried a cheaper one, but it failed to fit.  So I went with the Vanco 233306X 6" Right Angle Super Flex HDMI High Speed Male to Female Cable with Flat Top from Amazon.

I used standoffs to mount the PiDP-8/I, and then glued a 1/4 inch wood shim to the bottom of the box. This is to improve friction fit, serve as strain relief, and provide a surface to epoxy to.

I did some careful measuring and established a baseline for my connectors that would descend a bit into the wood shim.  Then I drilled a couple holes, and used a rotary tool and wood carver's tools to make the holes mostly shaped like the connectors.  A couple times, I lost control of the rotary tool and scraped up the box.  For example the line above the power connector you see in the final picture below. My holes are less than perfectly shaped.

The USB Hack:

The USB and hard-wired network connectors of the Raspberry Pie were quite close to the edge. Some people just cut a hole in the box, but I wanted to try and make an extender that would work in a confined space. But there is very little clearance between the Raspberry Pie and the case.

For me this meant cutting up a standard USB connector and hacking up a cable. My first attempt failed as the pins broke off. Here are the steps I took to get to the cut-down connector that did work.
Mouser Electronics was a quite reasonably priced source of male and female USB connectors.

1. Start with a USB Male connector.
2. Use a cutting wheel.
3. Cut the sides at the notch. 

4. Side cuts to guide the main cut.

5. Cut the metal.

6. Then the nylon 

7. Solder it up and your done?

8. The pins break right off.

9. Second try. Cut here.

10. Clear the nylon above the pins. 

11. The wires fit well in the well.

12. Power, Signal, and Chassis Ground.

13. Magic elixir!

14 Protects physically and electrically.

15. Carefully carve out  the case
where the USB connector will land.
16. NOW you are done!