DaedTech Digest: Hosting Woes and Unit Testing Studies

Happy Friday, folks.  Time for another installment of the DaedTech digest, wherein I link out to a bunch of technical posts I’ve written for other sites.

This week, I don’t have any specific personal narrative to relay, per se.  That’s largely because I’ve been working like a dog as we onboard additional clients to Hit Subscribe.  Oh, speaking of which…  we’re opening up the author rolls again, but looking for certain specific column types and topics.  If you think you might be interested, here’s the page where you can get started.

The only other item of note is that I have now successfully switched hosts.  My old hosting provider, Hostmonster, had my site on a server that must have become overloaded several weeks ago.  After seemingly endless support tickets and runaround from their tech support, they settled on three rotating and competing verdicts:

1. You’re right, our server is overloaded and we’ll fix it.
2. Actually, no, it’s probably WordPress or plugins or internet or something, whatevs, good luck, %#$& you.
3. Ah, actually, it is our server, but that’s life and you need to pay more if you want us to fix it.

I got tired of pleading my case for them to listen to the techs that said it was (1), so I gave them an ultimatum: migrate me to a new server, or buh-bye.  They made their choice, and so did I.  I’ll get to the new provider in a second in my picks, but suffice it to say that I’ve measured page load times on GTMetrix, and my identical, migrated site now performs substantially better than it even did on Hostmonster, even before they overloaded the server.  And I’m paying less!

I’ve installed no new plugins on my site or changed its configuration in months. But you’re probably right Hostmonster, this massive uptick in intermittent server errors and timeouts… probably because I wrote a long blog post or something. Couldn’t possibly be a resource issue on your server.

I’m planning to migrate a couple of my other sites as well in the coming weeks.

Picks

• On that note, my first pick is the new hosting company, Siteground.  It’s too early for me to speak to their long term performance or even support quality.  But, for someone who is too busy to deal with the mundanities of website administration, they are perfect.  I just had to give them my login credentials to Hostmonster, and they took care of everything — not just migrating the site over in its entirety, but taking care of the SSL cert, redirection, moving my DNS, and even taking over supporting my email.  I didn’t see any downtime, and I expended almost zero effort.
• Also related to this saga, I want to throw a nod out to isup.me, which is a great way to see if a website is down for everyone or if you’re just having a problem.  I took advantage of the “check my site for malware” option (to throw it at Hostmonster support, if they tried that excuse) and wound up chatting with the site owner.  He’s getting into hosting recommendations and other value-adds, so bookmark it and keep your eye on what they do.
• Speaking of things that abstract ugh IT issues away from me, I also pick LastPass.  I’ve used them for years now and it’s incredibly convenient.

The Digest

• I wrote a post for Scalyr about the idea of creating an audit trail for your business.
• For Typemock, I wrote a post about how to convince your management of the value of unit testing.  This is a subject near and dear to my heart.
• I wrote a post for the Rollout blog about how the future of getting more agile (in the adjective, more than the industrial agile complex, sense) involves increasingly granular releases.
• Here’s something I wrote on the Techtown Training blog, called about getting software developers more focused on business value.
• And, finally, a post for the NDepend blog.  This one generated a lot of buzz because I studied a hundred codebases and talked about the effects that more unit tests had on codebases.  I think it trended on Hacker News, so I cringe to think how many people there probably wished for my death.

And, on that note, have a great weekend!

DaedTech Digest: Delays, Risk, Visualization and Gigantic Log Files

Another week in the books, marked, as always, by a DaedTech digest post.  Another week here in sunny San Diego, marked by a predictable mix of work and fun.

On the work side, we’re building out our tech content agency at what feels like a fairly furious pace.  I suppose at some point, I may write DaedTech posts about business and growing an agency, but that seems a little off mission at the moment. (Weigh in if you feel otherwise.)  I’m also keeping the dream alive with our boutique codebase assessment practice, so a lot going on.

On the fun side, we spent a great weekend exploring Pacific Beach and La Jolla Beach in San Diego.  This included a ton of walking/hiking, sampling meals, appetizers, local beers, and desserts at local establishments, and just generally being outdoors.  As someone born and raised in the US Midwest, I can think of nothing better than a winter where it’s perpetually 70 degrees and sunny.

To include you in my joy, here’s a picture of some sea lions we came across while exploring.  Don’t let the apparent peacefulness fool you.  These things bark like crazy, stink like crazy, and spend most of their time banging their heads against one another.  It’s an endlessly amusing spectacle.

Picks

• In case you missed it, DaedTech has a new direction and I’m gaining steam with it.  Here’s a category tag for this series I’m going to be working a lot on that basically lays out a guide for how we, as software developers, can take over the software development industry, rather than serving as perpetual grist in its mill.
• I’ve been creakily easing my way back into software development, and Newtonsoft Json.NET provided a nice, gentle landing for not having to bother writing my own logic to parse JSON.
• Speaking of software development, you know about Alexa site rankings?  Well, if you want to get that data programatically, you might think you have to deal with AWS’s byzantine SOAP API, public and private keys and all that.  Nope.  Check this out.  That link is a neat little XML payload containing DaedTech’s Alexa rank.  Just swap whatever URL you’re curious about for daedtech.com in the URL parameter.
• On the personal front, I pick La Jolla beach, which is where we encountered both the sea lions pictured above and harbor seals.  If you’re ever in southern California and you like outdoors/nature, it’s worth a trip.

The Digest

• For Scalyr, I wrote a post describing the problem of having a log file that’s too big to open in an editor.  What should you do?  In case you couldn’t tell, the target audience for that is more on the newer dev side.
• On the TechTown blog, I did a post exploring how you calculate the cost of delays.  This is sort of inline with the Pluralsight course I did once about business cases for best practices.
• For Rollout.io, I explored the relationship between feature toggles and continuous delivery.
• This was a fun one.  Visualizing code is important, so I wrote a post for NDepend on tips for visualizing your code.
• I wrote a post for SubMain about low impact strategies for reducing risk in software development.
• And, finally, for the Hit Subscribe blog, I wrote a post about why your blog should be more like a wiki than like a newspaper column.

DaedTech Digest: Bad Coding Standards and the CRAP Metric

Another Friday, another DaedTech Digest, another recording of me spraying words all over the internet.  Fun times, as always.

Today (Thursday, at the time of writing), I went out to brunch.  This was sort of an accidental thing that happened.  We had someone come clean the condo we’re staying in here in San Diego, and it’s usually easier just to kind of make yourself scarce while that happens.  So we ambled out of our building and into downtown Ocean Beach, to kill time by eating breakfast.

Naturally, I worked while we did this.  My life is this kind of eternal blend of slow travel, work, and enjoyment.  So I sat, laptop open, drinking coffee, eating a “scramble,” and cleaning my inbox.  And, as I did this, it occurred to me how awesome the capability for remote work is.  It’s bled fully into my life over the last year, to the point where brunch on a Thursday is fine, and I take “weekends” when I feel like it.

Now, remote work might not be for everyone all the time.  But, in 2018, with the technology we have, it should be an option for you if you want it.  There’s a big, interesting world out there, and, armed with a laptop, there’s no reason for you to spend it in quiet cubicles or setting at noisy open office benches.

Picks

• That segues into my first pick, which is remote work.  If you have any inkling that you might want it, push hard for it, because it is worth it.
• Also, this is where I ate.  It was delicious, but, more interestingly, I described it to my wife as “UX epitomized in the meatspace.”  It was like a mid-2000s Apple device or a Cadillac.  No detail was too small to perfect about the food, presentation, decor, or, really, anything.
• Vlad Bown (with my permission) translated my first expert beginner post into Russian language.  Pretty cool!
• I’ve just got myself a copy of Typemock’s Isolator and plan to start using it for some code I’m writing for Hit Subscribe metrics gathering.

The Digest

• For Scalyr, I wrote a post describing, essentially, their core value proposition.  It was called “Log Management: What Is It And Why You Need It.”
• I continued my CodeIt.Right rules series for Submain.  This time, it covered “do not declare protected members in sealed types,” “avoid long type arguments for VB6 clients,” and “externally visible types and members should have XML comments.”
• I stirred the pot a little on the NDepend blog with a post about why code reuse is not a good goal.
• Another Scalyr post, this time about how to shorten the defect lifecycle.
• Here’s a fun discussion starter for SubMain: what’s the worst experience you’ve ever had with a coding standard?  Seriously, go weigh in — I know there are some good, facepalm stories waiting to be told in the comments.
• And, finally, I wrote about the ol’ CRAP metric.  Seriously.  That’s a thing, and not something that I made up.

DaedTech Digest: Proving That Singletons Hurt You and More

Another week, another DaedTech digest post.  It’s been a tiring, but productive week for me.

I did a Q&A and talk for the Developer on Fire Remote conference, had a whole bunch of meetings with our Hit Subscribe clients, and did a bit of R&D for my codebase assessment practice.  Oh, and on top of all of that, I tried here and there to get out a little and enjoy San Diego, where I’m spending the winter.  Oh, and I slept a little, but not much.

My wife, Amanda, and I have an extremely mobile, location-independent lifestyle.  For the most part, this is insanely awesome, and I’d recommend it for anyone.  But every now and then, the basics of life, like haircuts, healthcare providers, and pet logistics can get… interesting.  So sprinkle some of those logistics on top of an already-hectic week, and you get me, ready for a beer, a hammock, and a nap.

Picks

• Ultimello is a chrome plugin that gives you interesting capabilities for Trello, such as column card counts and the ability to sort in a lot of different ways.
• If you’ve never read it, check out Getting Things Done, by David Allen.  Strategies for managing your to-do list, inbox, and not feeling like you’re forgetting something.
• In San Diego’s Ocean Beach (OB), on Wednesday nights, they have the Ocean Beach Farmers Market, which features awesome produce, but also an unbelievable number of food and craft vendors with great stuff.

The Digest

• For the SubMain blog, I wrote a post to start a discussion about how people handle API documentation.
• I took an aggressive stance on dead code for the NDepend blog, saying there’s no excuse for it.  (This is doubly true when you have NDepend, since the tool helps you root it out).
• On the Scalyr blog, I made a post about ways that you can make things easier on the people reading your log files.
• Another one for SubMain.  This time I asserted that you don’t need code review templates.  Hint: just automate stuff instead.
• I once wrote a post for NDepend saying that singletons do bad things to your code.  People got worked up, so in this post, I proved it.
• And, finally, here’s another Scalyr post.  This one explains what log management is and why you need it (which is Scalyr’s bread and butter, incidentally).

DaedTech Digest: Serverless, Log Analysis ROI, and Some Static Analysis

Happy Friday, everybody, and welcome to another installment of the DaedTech digest.  Last week’s digest came after a frantic week of travel.  This week’s comes after an awesome week of San Diego.

I cannot overstate how much I think the future of knowledge work involves remoteness.  My wife and I run an agency where authors from all over the world produce content for clients all over the world.  But, beyond that, I’ve spent the week wandering around eating lobster tacos and jogging against this backdrop:

Your mileage may vary on how much transit you like and how much you might prefer settling in one place.  And you might really enjoy a lot of interpersonal human interaction.  But the allure of living and going where you want, when you want is powerful enough to shift the balance, I believe.

Anyway, I could evangelize for nomad lifestyle and remote work all day.  But instead, I’ll get to picks, and I won’t even pick remote work.

Picks

• If you’re interested in location-independent entrepreneurship (as a beginner or seasoned pro), I’d check out the Tropical MBA (TMBA) podcast.  They’re in the process of trying to re-brand since the name makes them sound like hucksters, but don’t let the name fool you — the content is seriously good.
•  Microsoft’s BizSpark program.  This is seriously awesome if you’re thinking of doing any kind of product thing.  You get 3 years of MSDN enterprise for free, provided that you’re actually serious about starting a product company (even as a side hustle) and that you want to use their software.  Their gambit?  That you start your business, get used to their stuff, and pay when you’re successful.  It’s a win-win.
• I feel like kind of a band-wagoning dope, but I’ll pick Slack this week.  It’s trendy and it’s ubiquitous, but it’s also indispensable for running Hit Subscribe, and I’m pretty pleased with it.  Also, they have the best 404 page ever.
• If you’re even in Ocean Beach, San Diego, go to the “Walking on Water Cafe” and get yourself a lobster taco.  There’s a pier that reaches out a long, long way into the ocean and, out near the end of it, there’s a little restaurant.  And they have lobster tacos for like $7.

DaedTech Digest

• I wrote a post for Scalyr about good application logging practices.  Keep your audience in mind, use a framework, provide context — that sort of thing.
• For NDepend, I wrote a post about cyclomatic complexity.  Well, kind of.  It was more about how you need more than just to measure average cyclomatic complexity to understand a codebase.
• I did another in my series of CodeIt.Right Rules, explained.  CodeIt.Right is a static analyzer that does automated code review, and I’m going through explaining the reasoning behind all of the rules.
• Another NDepend post, this one about the importance of visualizing your codebase with a dependency graph.  To really know what’s happening, forget whiteboards and design docs, and think automated tooling.
• And, another Scalyr post.  In this post, I talked about how one might go about calculating the ROI of a log analysis tool.
• And, finally, a post for Rollout where I talked about how you’d use feature flags with a serverless architecture.