Category Archives: Tech

E-publishing: Some Lessons Learned

Bolicks-Football-2014-v3-cover-400wide-229x300I just published (I hope) the Kindle edition of Bolick’s Guide to Fantasy Football Prospects 2014, JD Bolick’s survey of this year’s NFL rookies crop, handicapped for fantasy football players. That means it’s all Quarterbacks, Running Backs, Wide Receivers and Tight Ends.

I say “I hope” because the Kindle edition is still in review at Amazon. Sometime today they should approve it, more than two weeks after I was able to publish the iBooks and PDF editions. There are a few reasons for this, and since handy guides to epublishing are all over the internet (and available as eBooks), I thought I would add my two cents.

Publishing for the iPad with iBooks Author is easy.

Author works like a page layout program, and while it doesn’t appear to be very flexible, it is dead simple to enter text and graphics and convert them into an eBook. Bolick’s Football Guide has photos and stats tables. These were easy to enter and look on the page the way they’re supposed to, with little fussing necessary.

It is also easy to save that eBook as a PDF.


The only issue I had with Author for this project was the process for adding hyperlinks. This book has more than 500 hyperlinks to YouTube clips that illustrate nearly all of Bolick’s observations about each of the players covered. The process for creating a link in iBooks author is (assuming the Inspector is open): Copy link to the clipboard, highlight text to be linked, click the checkbox that says Enable Hyperlink, click dropdown and choose Link to Webpage, then paste in the link from the clipboard. This is three clicks too many, and those add up when you’re adding hundreds of links.

Publishing words for the Kindle is easy.

The Kindle started out as a reader. Load an eBook and it would let you read it on its screen. You could (and still can) specify the type size and style, and it will reformat the pages based on the user’s choices.

When I turned Bolick’s Guide to Fantasy Baseball Prospects 2014 into a Kindle book, after laying it out in iBooks Author, I just copied the text into a Word document. Then I saved the Word document as an htm file, and uploaded it to Amazon. It was easy.

That’s because creating an eBook with only words that the Kindle can read is very straight forward. You don’t have many options. And all of the guides available on the internet (plus Amazon’s own Guide) quickly tell you all you need to know. The most complicated thing is making the Table of Contents.

Publishing graphics for the Kindle is something else.

The Kindle is not a single device. There are the epaper black and white readers, and the color Fire ones. There are small screens and larger ones, and each comes with different expectations.

People who write about creating their eBooks, people like me, have their own experiences, and aren’t necessarily experts in all the different paths one might take to create a book. I’m no expert, but I offer up the following just in case you’re hoping to create a Kindle book with a lot of tables in it.

Amazon discourages you from using tables in your layout to organize different parts of the page. Using tables to create html layouts was a common hack in web page design in the days before css. It was frowned upon, but also sometimes the easiest way to get things to line up the way you wanted.

Using tables to similarly hack page design for the Kindle is a mistake, but Amazon isn’t as clear in all places that tables full of tabular information are also not welcome.

In the Guide we use charts like this one, to show a player’s statistical history in college.

Screen Shot 2014-08-18 at 10.55.43 PM

When I converted the iBook version into a Word document, the tables were converted into html tables via the TABLE tag. They displayed fine in Word and later in Dreamweaver, when I got away from using Word, but were a jumble when I viewed the file in the Kindle Previewer or in Chrome.

I wasted all sorts of time trying to change the settings. I noted that someone, maybe Amazon, said that tables wouldn’t display properly in Kindle Previewer, but I still couldn’t get them to display in Chrome. I tried to upload my layout full of tables to KDP, the publishing arm of Amazon, and it was rejected. It was only then that I came across a caveat that tables were not supported for the Kindle at all. It was suggested I use jpegs instead.

I spent yesterday making screen grabs of all the stats boxes in the iBook version and replacing the tables in the Kindle version. When I was finished I tried publishing again. Success.



Netflix Fail Confirmation? Not so simple, but maybe

Some months ago I wrote about our miserable experience with Netflix. Constant freaking buffering through our admittedly old Roku box.

We tried a number of things to improve performance, including a new N-type router, but the same behavior persisted until we learned to go with the flow, sort of.

Instead of reloading the network, or reloading the stream, the best thing to do was wait: We would start a stream, it would play for a few minutes, grind to a halt, then reload and eventually drop from 4 dots of throughtput to 2 stars. It would then buffer some more and after a couple of minutes, it would work. Usually.

Sylvester Stallone would be grainier. So would Denise Richards. But the show would go on.

This all started happening late last year. At first I thought it was perhaps interference from all the routers in our building and the ones surrounding it. Or a microwave or cordless phone on the other side of the wall.

But the more I diagnosed the cause, it seemed to all come back to the fact that Netflix used to seamlessly adjust the screen resolution based on the amount of available bandwidth, but it was now insisting on running in HD. Despite the fact that Verizon cannot deliver more than 3Mb per second to our house over DSL.

Which is fine. I get HD when I stream on my computer. Barely. But the Roku fails and that seems to be a failure of either Netflix to properly adjust the resolution or Roku to transmit the information.

In any case, as much as I loathe Verizon, they didn’t seem to be particularly to blame here. My bandwidth tests show occasionally erratic bandwidth, but more often not delivery of what I was paying for. And my stream, while far from faultless, wasn’t constantly buffering. It was, in fact, constantly adjusting stream rates so I had a picture. Sometimes really sharp, sometimes pixilated. Usually with inane local banter.

I wrote about this here some months ago (February), mainly my attempt to work out what was going on. And it seemed to make the most sense that Netflix was trying to force the ISPs to provide more bandwidth (Netflix consumers 70 percent of the streaming video volume, and insane amount of total internet use every day) at lower cost. Those that did, had reasonable throughput. Those that wanted to charge Netflix for that throughput, namely Comcast and Verizon, saw throughput drop.

In recent weeks, Netflix started to send an error message to customers saying that the buffering was Verizon’s fault. This caused Verizon to go to court, and before it showed any evidence Netflix withdrew the error message.

Which makes this rather long and somewhat technical story of interest, since it seems to muster numbers that show that the Comcast and Verizon problems for Netflix were actually caused by Netflix. Sort of like I said in February.

As the story concludes, it is based on the available numbers, and those may not be all the numbers.  Netflix may actually have good reasons to be negotiating with Comcast and Verizon this way. And it seems likely that there will eventually be a better way to handle the sharing of routes of pipes, called “peerage,” in the future.

For now, we’re stuck with this mess. And I don’t think it unreasonable for Netflix to try to push Comcast and Verizon into better service. These monopolistic giants are a burden to us all, and we would be better off with a different system.

Verizon signed a franchise agreement with the city of New York in 2007 that said it would provide universal fiber optic (FIOS) service throughout New York City by the end of June 2014. Uh oh.

It turns out that FIOS installation is really expensive. Verizon got slammed by our two big storms, Irene and Sandy, but the fact is they’re giving up on FIOS. Having no competition, they can sit on their decaying DSL system, invest less and milk their exclusive franchise until someone sues them for failure to deliver. Which will mean a fine, a cost of doing business, in the future, and leave large parts of New York City a tech black hole, and yet a cash cow, for the phone company.

I want my video now. And if Netflix is messing up my stream, they’re playing with fire. The problem is they have the extinguisher: No video.




Verizon May Suck A Lot, Or Only A Little Less Than A Lot. But I Think Netflix is the Problem.

This story at arstechnica lays out some ambiguous broadband stats from Netflix about various ISP’s deliverance of our video feeds to our house.

In my house we’ve noticed in recent months that we can rarely watch Netflix, delivered through our first generation Roku, even after I moved the router into the same room (about 20 feet apart).

And that comes after years of perfectly fine service. Somehow the system is getting worse.

At first, the chart in the article suggests that Verizon (and Comcast) are throttling performance, but the evidence for that doesn’t seem to exist (though Verizon’s recent net neutrality victory is grist for the throttling mill, and a warning of what could happen if our internet pipes aren’t protected from pipe-holder taxation.)

What I know for sure is that Netflix relentlessly tries to deliver a HD signal into my house. My HD TV loves that, but my contract with Verizon is for a fairly modest bandwidth (3mb down, 1mb up, the max their system can somewhat reliably deliver). Whenever we watch Netflix, we have to set the program up and then wait either a long or an interminable amount of time for Netflix to figure out that we don’t have the throughput to handle the signal they want us to have. Once we go from HD four dots to two dots, based on their evaluation, we can watch our TV, usually without problem.

But this transition always takes a stupid amount of time. WE DON’T HAVE THE BANDWIDTH FOR HD, we scream, but Netflix can spend scores of minutes trying to pump the HD our way. And does not seem to memorize our settings, nor allow us to set our own (gimme gimme gimme two dots!)

Tonight we waited nearly a half hour (doing other things, too, we’re not hopeless) waiting for Netflix to tamp down our usage rate so we could watch our show, and then quit because it didn’t happen.

Netflix used to tamp down bandwidth rates with great agility. I’ve read articles about how they maximized flexibility, and valued their ability to reduce their bandwidth footprint, but that no longer seems to be the case. I want to blame Verizon for this, since they offer fairly crappy service on my block, but I think the greater problem is that Netflix for some reason no longer values that elastic delivery.

They want to deliver HD even if you’re not capable of receiving it, and that’s screwing up my watching of Season 5 of Breaking Bad. I dislike Verizon, but it seems that Netflix is the one who can fix this problem.

Update: It looks like Netflix agrees the problem isn’t Verizon.

Update (February 22): It looks like Ars Technica now thinks the problem is Verizon, demanding substantial peering payments.

New Sherlocks Start Tomorrow! Yawn.

what-calm-i-m-waiting-for-sherlock-season-3-6But not for the reason you think. My daughter was so excited by the prospect of new episodes of Sherlock that I hooked up Tunnelbear to my internet connection. Tunnelbear is VPN software that anonymizes your data stream, so you can watch local content from other countries. So we streamed Sherlock from the BBC1 website in the UK, just as if we were holed up in our seaside cottage in Port Wen in Cornwall, eating pasties.

Tunnelbear works great, and it also encrypts all your data, so you can’t be tapped by evil doers in public wifi hotspots and the NSA everywhere. At least in theory. They gift you with 500mb of data per month, and if you Tweet their praises @thetunnelbear they’ll give you another GB, which combined is enough for more than an hour of TV watching a month from England or France or a few other countries they have servers. For $5 a month or $50 per year you can use the service as much as you like. This advertisement generates no remuneration, it is just an expression of amazement and enthusiasm.

With a one-month subscription our little one and her little friend have seen the whole season of the gangly detective and his very nice sidekick, already, and won’t be troubled by the late night scheduling on a school night during mid-term time that caused so much consternation last time. No spoilers, but I’m told episode 2 is just the best Sherlock ever.