Yesterday, I did another speed test with CLEAR‘s level 1 tech support. (I.e., the reps you get by dialing 1-888-888-3113.)

CLEAR’s support rep said they had done no work on my ticket, but they wanted to do another speed test anyway. Hrm. No work at all? None.

The results: My bandwidth is now about 5Mbit down, 1Mbit up. I no longer have a basis for a complaint, so I asked them to close my problem ticket.

Why is it now 5Mb/1MB, when three weeks ago it was 1/10^th that? There’s no explanation. CLEAR claims they did no repairs that would have affected my connection. And I didn’t do anything here to affect it. Yet it’s much faster. It’s a mystery of God’s creation.

Does this happy if unexplained ending let CLEAR off the hook? No. Their customer & technical support were atrocious. They were never straight with me about when they’d work on my problem — indeed, they’re now saying they did no work! And regardless, my connection bandwidth is still below their advertisements, but not sufficiently bad so that I’d complain about it. Which, I’m sure, their financial bean counters are counting on.

For the record, Qwest customer and tech support is 1,000,000,000,000 times better than CLEAR’s, and their DSL product is rock solid. “If I knew then what I know now,” I wouldn’t have switched. But it’s finally working satisfactorily, so I’m going to leave well enough alone.

CLEAR contacted me. There’s no new information about my problem, and they didn’t say they fixed anything. But they want to do another speed test.

I’ve scheduled it for tomorrow evening.

Uh-oh

I quickly noticed a time lag in my surfing. Speedtest confirmed that the bandwidth was not what CLEAR had advertised.

CLEAR had promised me 7Mb/1.2Mb when I signed up. I was seeing 650Kb/30Kb, at best. I’d have faster bandwidth by etching bits onto rocks and throwing them at passing cars.

I verified all my connections, and tried swapping cables. The modem was near a window, and its connection indicator displayed three or four LEDs out of five.

I checked CLEAR’s unhelpful online help. “Position your modem near a window.” Check. “Make sure there’s no other equipment nearby.” Check. “Clear your browser cache.” Give me a break, but check. Still poor bandwidth.

The “chat help” Horror

I then turned to their site’s Support page, which directed me to use their online “chat” help.

The online “chat” support reps are measured by how well they prevent a customer from escalating to the next level of tech support. This is not not not identical to solving customers’ problems! They want to prevent you from speaking to a competent live person on the phone who doesn’t have his or her head embedded up their rectum knows their stuff, because it costs CLEAR more money to talk to you than to “chat” with you. Either (gasp!) solving your problem, or wearing you down the point where you give up, are the same, to CLEAR’s short-term financial position.

To make you comfortable, the reps use your first name in nearly all their responses. I found this imbued the exchange with a cloying false friendliness. They were obviously selecting canned responses from a software system that inserted my first name into a variable. “I am happy to try to assist you, John.” “Please clear your browser cache, John.” “Please look at these naked pictures I have of your sister, John.” Etc.

And WTF is it with online services asking you to clear your browser cache?! Holy Mother of God. Clearing my cache has helped in maybe .0025% of all the times I’ve had browser problems in my entire life. When I read a published paper showing that browser caches slow down the Internet, I’ll take the “Please clear your browser cache” requests seriously. Until then, please fuck off.

The “chat” help was unable to help. Their final advice was to try powering everything off, and then on. I am not kidding.

Horror #2

The next day, I decided to call their tech support desk and talk to a real person. Their web site had a pulldown menu for phone numbers, which didn’t work. Hmmm.

So I started another chat session to ask for their tech support number. The rep was obviously loath to give out phone numbers. I had to threaten to cancel my CLEAR contract to get him to cough up a number. Here’s the exchange:

I do not want to help you, John.

I called the number, and this rep asked me to remove my AirPort Extreme from the equation. The AirPort had worked fine when I was on Qwest 7Mbit DSL service, but I did as she asked. The performance was still abysmal. I called back, and was phone rep asked me to reboot my Mac in Safe Mode. Still bad performance. After quite a few more speed tests, I was escalated to CLEAR’s third level of customer support, their “Case Management Team.”

Good Stuff, Briefly

CLEAR’s Case Management Team rocks! They were thorough and very knowledgeable, with a minimum of BS!

More speed tests, a traceroute, and a remote analysis of the nearest WiMAX antenna. A repair ticket was entered for the problem. The rep didn’t commit to anything, but he suspected the problem was within their network. I was given a case number, and told I’d hear something within 24 to 48 hours.

I was happy.

Horror #3

Forty eight hours later, no call.

I called back. There had been no activity on my ticket, and I should call back if I hadn’t heard anything within three more days.

Horror #4

I waited four days, and called again. (We’re now six days out from the problem ticket’s creation.)

This time they said there was no ETA for the ticket! No action had yet been taken on it, and there was no guarantee, estimate, or guideline that I would hear anything within any amount of time. Awesome!

I went home that night and discussed this with my spouse. We decided to give CLEAR more time.

I called back the next day to ask for a refund on my first week of CLEAR. Gee whiz, that got someone’s attention. This rep now told me I’d hear something within seven to 10 days.

24 – 48 hours! Three days! Never! Seven to 10 days! How about I shove the CLEAR modem up my posterior orifice and sing Zip-A-Dee-Doo-Dah?

I’ll post the next installment of this saga in a few days, when the “7 – 10 day” period expires. Care to place bets?

I verified all my connections, and tried swapping cables. The modem was near a window, and its connection indicator displayed three or four LEDs out of five.

I checked CLEAR’s unhelpful online help. “Position your modem near a window.” Check. “Make sure there’s no other equipment nearby.” Check. “Clear your browser cache.” Give me a break, but check. Still poor bandwidth.

The “chat help” Horror

I then turned to their site’s Support page, which directed me to use their online “chat” help.

The online “chat” support reps are measured by how well they prevent a customer from escalating to the next level of tech support. This is not not not identical to solving customers’ problems! They want to prevent you from speaking to a competent live person on the phone who doesn’t have his or her head embedded up their rectum knows their stuff, because it costs CLEAR more money to talk to you than to “chat” with you. Either (gasp!) solving your problem, or wearing you down the point where you give up, are the same, to CLEAR’s short-term financial position.

To make you comfortable, the reps use your first name in nearly all their responses. I found this imbued the exchange with a cloying false friendliness. They were obviously selecting canned responses from a software system that inserted my first name into a variable. “I am happy to try to assist you, John.” “Please clear your browser cache, John.” “Please look at these naked pictures I have of your sister, John.” Etc.

And WTF is it with online services asking you to clear your browser cache?! Holy Mother of God. Clearing my cache has helped in maybe .0025% of all the times I’ve had browser problems in my entire life. When I read a published paper showing that browser caches slow down the Internet, I’ll take the “Please clear your browser cache” requests seriously. Until then, please fuck off.

The “chat” help was unable to help. Their final advice was to try powering everything off, and then on. I am not kidding.

Horror #2

The next day, I decided to call their tech support desk and talk to a real person. Their web site had a pulldown menu for phone numbers, which didn’t work. Hmmm.

So I started another chat session to ask for their tech support number. The rep was obviously loath to give out phone numbers. I had to threaten to cancel my CLEAR contract to get him to cough up a number. Here’s the exchange:

I do not want to help you, John.

I called the number, and this rep asked me to remove my Airport Extreme from the equation. The Airport had worked fine when I was on Qwest 7Mbit DSL service, but I did as she asked. The performance was still abysmal. I called back, and was phone rep asked me to reboot my Mac in Safe Mode. Still bad performance. After quite a few more speed tests, I was escalated to CLEAR’s third level of customer support, their “Case Management Team.”

Good Stuff, Briefly

CLEAR’s Case Management Team rocks! They were thorough and very knowledgeable, with a minimum of BS!

More speed tests, a traceroute, and a remote analysis of the nearest WiMAX antenna. A repair ticket was entered for the problem. The rep didn’t commit to anything, but he suspected the problem was within their network. I was given a case number, and told that I’d hear something within 24 to 48 hours.

I was happy.

Horror #3

Forty eight hours later, no call.

I called back. There had been no activity on my ticket, and I should call back if I hadn’t heard anything within three more days.

Horror #4

I waited four days, and called again. (We’re now six days out from the problem ticket’s creation.)

This time they said there was no ETA for the ticket! No action had yet been taken on it, and there was no guarantee, estimate, or guideline that I would hear anything within any amount of time. Awesome!

I went home that night and discussed this with my spouse. We decided to give CLEAR more time.

I called back the next day to ask for a refund on my first week of CLEAR. Gee whiz, that got someone’s attention. This rep now told me I’d hear something within seven to 10 days.

24 – 48 hours! Three days! Never! Seven to 10 days! How about I shove the CLEAR modem up my posterior orifice and sing Zip-A-Dee-Doo-Dah?

I’ll post the next installment of this saga in a few days, when the “7 – 10 day” period expires. Care to place bets?

A man may choose to tinker with something that’s not broken, and look for “better” alternatives. I’ve done that to my Internet access. Woe is me.

I’ve occasionally thought about switching to home WiMAX. The reasons include mobility, if we add on mobile service; fewer phone cords in the house; more latitude in configuring our home offices; and maintaining our Internet access if we move. And so last week, I made the switch: I cancelled Qwest, went to the CLEAR website, and signed up for their “Home Internet” premium service. It’s $40/month, and promised 7Mbit down (or 6Mbit, depending on the sales document) and 1Mbit up. All the stock image photos had smiling faces. What could go wrong?

My experience with it has been terrible. So much has gone wrong that I’ll have to spread the bad news across multiple posts.

Let’s talk about signing up, setting up, and the Terms of Service.

Signing up

Hmmmm…. Signing up was easy and fast. These aren’t the droids you’re looking for, move along.

Unpacking and setting up

The equipment has a sleek design. Setting it up was a little confusing, however.

They ship you Motorola’s “read me first” flier, as well as their own “read me first” flier. So…which one should you read first?

I chose to read Motorola’s RMF first. It tells you to surf to a 192.168 address to configure the modem. Simple! Alas, it doesn’t work, and never can work, because nowhere in the box or paperwork will you find the modem’s password.

A call to their customer service later, and I learn that I should completely ignore Motorola’s RMF. Hey CLEAR here’s an idea: Maybe you shouldn’t ship that piece of paper to your customers.

I’m told to just surf to any web address and the modem will automatically redirect me to CLEAR’s page for completing my account registration. And, yes, it works and I continue with my account registration.

And then…

The ToS Horror

It was at this point that I first got my inkling that my love affair with CLEAR would be short.

The Terms of Service is a horror show.

Their marketing paraphernalia doesn’t tell you ahead of time, and no salesperson will volunteer this: To get CLEAR service, you must agree to receive any and all spam on your phone number, regardless of cost.

Let me repeat that: To get CLEAR service, you must agree to any and all spam on your phone number. And agree to pay any mobile charges for it. Think I’m kidding? Here’s the agreement (click to enlarge):

CLEAR Terms of Service

Regardless of whether your phone number is in the national Do-Not-Call Registry, you consent to CLEAR and any designated agent calling you for any purpose and by any means. You must agree to receive any number of calls for any reason from any party. And if this costs you money, due to your mobile carrier’s account terms, that’s fine by you.

Did I agree to this? Yes. I had already cancelled my prior service. (Yes, I was too trusting stupid in doing so.) Umpteen other items in my life were demanding my attention, I didn’t relish having to mail back the equipment, I was looking forward to getting on the net, and I just wanted to get on with it. So I held my nose and agreed to their ToS.

If I had known about it when considering whether to switch, I wouldn’t have signed up. But…I’m 99% of the way there. So yes, I did.

CLEAR, I’m sure, made a calculated decision to tell their prospective customers about this at the proverbial last moment. They’re expecting most customers will have the same reaction I did: “Oh, you assholes. Now you tell me. [expletive string] OK, fine.”

Once I completed this step, I was on the net, with astounding bad performance, to be followed up by atrocious customer support. That’s a story for the next blog post.

My life however is in a bit of upheaval, because we’re refinishing half of our home’s wood floors.

stairs...

study...

stain choices...

gaaaaa!

Seattle Public Library

Queen Anne Ave N

Queen Anne water tower

Office desk

Today’s Undocumented Project Hall of Shame Exhibit A is Sunburnt. Its author describes its rationale in a blog post that practically had me salivating. I want to replace Lucene with Solr, but don’t want to leverage Django’s ORM to do it. Although Sunburnt sounded like the solution, its total documentation turned out to be a ~25-line readme. Good grief. It might be the greatest thing since sliced bread, but I won’t touch it without docs that include a working example.

“Wait, just read the code!” Nuh-uh. Slippery slope, not walk down I.

If you’re giving open-source code to the world, do the right thing and document it. Any project into which you invest your energy, and of which you think highly enough to publish as open-source, is worth the time. Conversely, if you can’t be bothered to document it, it’s crap that’s not worth publishing.

SEO “expert” madness! Can I get paid good money to give companies shitty advice?

Beautiful is better than ugly.
Explicit is better than implicit.
Simple is better than complex.
Complex is better than complicated.
Flat is better than nested.
Sparse is better than dense.
Readability counts.
Special cases aren’t special enough to break the rules.
Although practicality beats purity.
Errors should never pass silently.
Unless explicitly silenced.
In the face of ambiguity, refuse the temptation to guess.
There should be one– and preferably only one –obvious way to do it.
Although that way may not be obvious at first unless you’re Dutch.
Now is better than never.
Although never is often better than *right* now.
If the implementation is hard to explain, it’s a bad idea.
If the implementation is easy to explain, it may be a good idea.
Namespaces are one honking great idea — let’s do more of those!

Out of all my T-shirts, it and one other have gotten the most comments. In the past month… A woman asked me if my shirt was about a religion, after reading just the back. I briefly explained it was about a programming language, and she seemed baffled. And, a man told me he enjoyed reading it, even though he didn’t know what Python was, and thanked me for the good thoughts for the day. And, I overheard two people remarking about it as I walked out of a cafe, talking about how interesting the philosophy was.

What other computer language has tenets that inspire these kinds of reactions from laypeople?

I’ve done my share of Java development in the distant past, and I can’t imagine ever going back. The static environment, names like Field.Store.YES, the layers upon layers of syntactic and semantic sludge…

Hmm, what would be a “Zen of Java”?

Turgid verbosity is better than brevity.
Multiple lines of code are better than concise simplicity.
Mid-1980′s software techniques are better than modern adaptable technology.
Many legacy libraries are better than clean organization.
Slow, costly development with large overhead is better than agility.
It’s better to type lots of soul-deadening crap having the grace of four-day old road kill than to quickly solve a problem.

The quick details

Chapter 1: What’s Django and why it’s good for e-commerce. Zzzzzzzz…

Chapter 2: Quickly getting going on your e-commerce project. Good intro.

Chapter 3: We’re getting some meat now. The authorization model, user profiles, django-registration, shopping carts.

Chapter 4: Payment processors, using Google, Amazon, and PayPal as examples. If I was the author, I wouldn’t have included Google. But otherwise, good stuff. The narratives that walk you through alternatives and subtleties are just as useful as the code samples, if not more so.

Chapter 5: Backend stuff. Good.

Chapter 6: Searching through the product catalog. Eh… Search is a mighty big topic. If your e-commerce site is small, you don’t need this chapter, because you can use just the ORM, or PyLucene if you want to go hog-wild. If your site is large, this chapter will only start to acquaint you with search issues, and you’ll need more information to do a good job.

I did note a couple of code sample oddities. E.g.,

if model is None:

instead of

if not model:

Chapter 7: Report generation. It’s enough to get you started, but you’ll have to do more reading and experimentation to get the job done.

Chapter 8: JavaScript and Ajax, using YUI and jQuery as framework examples. I wouldn’t have included YUI, but that’s me. Like the catalog-search chapter, this will illustrate what you’ll confront, but it’s nowhere near enough to get the job done.  To put it in perspective, the chapter is only 21 pages. If JavaScript and Ajax is a lake, 21 pages is just enough to post the “Danger! Thin ice!” sign.

Chapter 9 is about selling downloadable products and services. Although well-intentioned, like the JavaScript chapter, there’s not enough information to do the job on a real commercial site. There’s about 20 pages on integrating with AWS, a subject about which some consultants base their entire business. The last time I worked with AWS I nearly threw myself out a window. It’s hideously complicated. At best, this will whet your appetite, and you better understand that you’ll need a lot more information and alcohol to do the job.

Chapter 10 is about deployment. Fabric, buildout, apache2, mod_wsgi, etc. Decent.

The quick summary

B-. If you’re a novice, you’ll get the lay of the e-commerce land. But you’ll have to then find other, more detailed, information sources. If you’re an intermediate or advanced developer, you’ll breeze through this in one sitting, and probably pick up a couple of useful nuggets of information. I wouldn’t rush out to buy it, but it’s a decent book for your shelf.

I didn’t give back as much as I wanted to. Well, that’s a little dishonest – I don’t think I gave back at all. I intended to (I had a crackin’ idea for a lightning talk) but I didn’t carry through. My bad. I will do better next time.

I sent Selena some comments about it:

I tremendously enjoyed this year’s conference. I attend only two conferences each year, PyCon and OSB, and I find the energy level and attitude similar and equally enjoyable between them.

I may not attend your wrap-up session today at 3:30. So I’d like to send you this mail with my feedback. I hope you don’t mind.

Please take this only as constructive criticism, to try to help in a small way. In doing this, I FULLY recognize that I might not be your target demographic or psychograhic! How much to factor in any of the following is up to you.

Feedback and suggestions:

1. This year’s location was better than last year’s. It’s good to actually be in downtown PDX. But I’m not saying that last year’s location was awful! If you could save money by going back to OCC or some equivalent place, I would consider doing that.

2. Although the external location was better, the internal room logistics were much worse this year.

a) The curtain “rooms” that weren’t rooms were terrible. It was too easy to be distracted by other presentations, or by hallway (fabricway?) conversations. A conference’s number one goal is to transfer information! The fabric “rooms” worked directly against that.

b) “Hallway” space is needed. In the OCC last year, hallway conversations happened in the, well, hallways. They were a venue for information exchange. But this year there wasn’t a hallway. There was a stairway, up and down which the conference attendees trudged. The dynamic was completely different. Eyes front, look down, don’t trip fall and die.

c) The hacker lounge didn’t feel like a lounge. It felt like a bunch of tables in the back of a room. The energy at least year’s lounge at the Hilton was much better.

3. There ought to be some thought given to additional quality control for next year’s talks.

a) I attended three talks that surprised me by how much they aimed at beginners. The drawback to narrative descriptions is that, while they allow for detail, they also allow for varying interpretations. “We’ll start with…. and then talk about….” can be interpreted in many ways. There needs to be some easy-to-scan indicator of a talk’s level – newbie, intermediate, etc.

b) Two talks veered dangerously close to being glorified marketing spiels for the products of their speakers’ employers. Even if not deliberately intended, it’s easy for biases to creep into a talk, or to give the appearance of a conflict of interest. I suggest that OSB adopt a rule along the lines of, you can’t give a talk about a product sold by your employer. Or, if your talk covers your employer’s product, you have to have a co-presenter who works for a different company.

c) There should be more formalized feedback to the speakers, and to OSB, beyond a wiki page on which anyone can change or obliterate others’ comments. Why not use Surveymonkey or something similar?

4. I know you have to balance money vs. goodies, but, the coffee/tea serving during the day was weak. More cookies.

5. I thought of a sponsorship prize for next year: Have your name used for a session room! (“The DeRosa room…”)

6. I liked the smaller paper program, compared to last year. You could save even more money by not having paper programs at all. Everyone’s lugging a laptop. Paper is so 20th century.

7. Some of the speakers were better than others at staying on time. Consider having session runners give the speakers 5-minute warnings, supervise audience questions, etc. Alternatively….give speakers the same instructions you gave them this year, but be more emphatic. One talk ran nearly 15 minutes over, and that’s a problem for the audience. Some of us have no problem standing up and walking out. Others are more timid.

14:30: NSFW. Don’t remember what this is, but the name intrigues. Heh, among other things, fun with Chat Roulette.

14:15: It was.

13:30: Advanced Trolling. This should be good.

12:06: Lunch…

10:15: UX and usability discussion.

9:04: Selena puts up slide saying this is the FINAL DAY. Gah! Not of all reality!

9:01: Getting under way. Cap’in, the engines donna work and there’s nothin ah kun duu aboo’ it.

8:46: Today’s the unconference day. I’m dragging. Warp engines are off-line. I’m still rebooting this morning. Lights on but nobody’s home. Pick your favorite analogy.