I updated the design of my blog over the weekend.

My main goal was to switch back to a new, responsive version of my original blog theme, with its orange and blue colour scheme. Since the start of this year I’d been trying out a variety of off-the-shelf WordPress themes to give me a responsive (and therefore more SEO-friendly) design, but I was never really satisfied with any of them, so orange and blue is back, with its first major design refresh since 2011.

I’d thought that making it responsive would be a massive undertaking, but in the end of the day it only took me a couple of hours on Saturday morning. I was helped greatly in this by the fact that I was already using Less CSS as a pre-processor. I was also able to use Google’s mobile friendly test tool to quickly identify and fix the issues that needed fixing.

If you shrink the window down to below 400 pixels, you’ll see that that the text in the header starts to shrink with it to fit. I’d seen a few other sites do this, and it turns out to be very simple to implement, using a combination of media queries and viewport units. In case you’re interested, here is the Less CSS mixin that I’m using to achieve this:

.responsive-font-size(@size, @resize-below) {
    font-size: @size;
    @media screen and (max-width: @resize-below) {
        font-size: unit(@size * 100 / @resize-below, vw);
    }
}

I’ve also restored my old blog posts, together with most of the comments and attached pictures, from a backup that I’d forgotten that I had. In the process of doing so, I’ve updated all the internal hyperlinks and image locations to point to https:// URLs on my blog; this was needed to eliminate mixed-content security warnings in the browser address bar on older posts that contained images. I achieved this quite simply and elegantly by using sed on the output of mysqldump to perform an appropriate find and replace before reloading. The exact command to use is left as an exercise for the reader.

Behind the scenes, I’ve moved it onto a $5/month 512MB DigitalOcean droplet. This is all you need for a blog that only gets a hundred or so hits a day, and a bit of load testing with Apache JMeter suggested to me that it should be able to handle a spike from Hacker News if necessary — apparently hitting the HN home page can get you about 6,000 hits an hour. I’ve scripted the server setup using Terraform and I’ve also got a couple of scripts to backup and restore the data. This means that I can tear down and rebuild it very quickly if need be, in accordance with the modern best practice of treating your servers as cattle rather than pets.

If you have any problems with it, or if anything doesn’t look right, please let me know. If you want to be reminded what the old version looked like, here are some archive.org snapshots for 2007, 2010, 2011, 2012, and 2016.

Posted at 08:00 on 27 November 2017.
Tagged: blog, css, design

I received an e-mail from a colleague the other day about some code that I’d recently pushed to GitHub. Since I’d pushed some more changes round about the time he sent the e-mail, I needed to know which revision he was referring to.

There’s just one problem:

Of course, I could have got the proper times from SourceTree or typing git log in the console, but it’s still annoying, especially since the GitHub page was more easily to hand. And GitHub does show you the exact time as a tooltip—something I missed at the time—but it’s still annoying, especially if you have to hover over half a dozen different datestamps to find the one you’re looking for.

We need to have a rethink about fuzzy dates. Yes, I know that it’s friendly and cuddly and warm and fuzzy and cute (and more interesting to code) to say “two days ago” or “eighteen hours ago,” but when I’m trying to refer back to 17:22 precisely, it’s utterly useless and just adds friction without providing any value whatsoever.

Posted at 08:00 on 29 February 2016.
Tagged: dates, design, ui/ux

Separation of concerns is often cited as the reasoning behind the traditional three-layer architecture. It is important, otherwise you will end up with a Big Ball of Mud.

However, in order to separate out your concerns, you must first categorise them correctly as either business concerns, presentational concerns or data access concerns. Otherwise you will end up with unnecessary complexity, poor performance, anaemic layers, and/or poor testability.

Unfortunately, most three-layer applications completely fail to categorise their concerns correctly. More often than not this is because it is simply not possible to do so, as some concerns fall into more than one category and can’t be refactored out without introducing adverse effects. I propose the term inseparable concerns for such cases.

The key to separation of concerns is to let it be driven by your tests. Under TDD, the first thing you would do if a particular line of code contained a bug would be to write a failing unit test that would pass given the expected correct behaviour. It is what this test does that tells you whether the code under test is a business concern, a data access concern, or a presentational concern.

It is a presentational concern if the test simulates raw user input, or examines final rendered output. For example, mocking any part of a raw HTTP request (GET or POST arguments, cookies, HTTP headers, and so on), verifying the returned HTTP status code, or examining the output generated by a view. In general, if it’s your controllers or your views that you’re testing, it’s a presentational concern.

It is a business concern if the test verifies the correctness of a business rule. Basically, this means that queries are business concerns, period. If they are not returning the correct results, then they have not implemented some business rule or other correctly. Other examples of business concerns include validation, verifying that the data passed to the database or a web service from a command is correct, or confirming that the correct exception is thrown in response to various failure modes.

It is a data access concern if the test requires the code to hit the database. Note that this is where the so-called “best practice” that your unit tests should never hit the database breaks down: if you are adhering to it strictly, sooner or later you will encounter a bug where it stops you from writing a failing test. Most people, when confronted with such cases, skip this step. Don’t: TDD should take precedence. Set up a test database and write the test already.

It is an inseparable concern if it falls into more than one of the above categories. Pretty much any performance-related optimisation that you do will be an example here. For example, if you have to bypass Entity Framework and drop down to raw SQL, you will have to hit the database to verify that business logic is correct. Therefore, it is both a business concern and a data access concern.

Inseparable concerns are much more prevalent than you might expect. IQueryable<T> is the best that we’ve got in terms of making your business and data access layers separable, but, as Mark Seemann points out, it still falls short because NotSupportedException. Another example is calling .Include() on a DbSet to include child entities. Although this is a no-op on Mock<IDbSet<T>>, you can’t verify that you are making the correct calls to .Include() in the first place without hitting the database. Besides which, if you’re mocking DbSet<T> instead of IDbSet<T>, as you’re supposed to be able to do with EF6, calling .Include() throws an exception.

I would just like to stress here that inseparable concerns are not an antipattern—they are a fact of life. All but the simplest of code bases will have them somewhere. The real antipattern is not introducing them, but trying to treat them as if they were something that they’re not.

Posted at 08:00 on 16 February 2015.
Tagged: n-tier deconstructed, tdd, design

Most developers use the term “Repository” to refer to a wrapper or abstraction layer around your O/R mapper, supposedly to let you switch out one persistence mechanism for another. However, if you look at its definition in its historical context, you’ll see that this isn’t what it refers to at all.

The Repository pattern is a part of your O/R mapper itself.

The Repository pattern was first described as follows in Martin Fowler’s Patterns of Enterprise Application Architecture:

Mediates between the domain and data mapping layers using a collection-like interface for accessing domain objects.

Patterns of Enterprise Application Architecture was written in 2003, at a time when O/R mapping technology was in its infancy. Most ORMs were commercial products, very simple by today’s standards — more akin to the likes of Dapper or PetaPoco than to modern heavyweights like NHibernate or Entity Framework. Hand-rolled data access layers were very much the order of the day. Furthermore, many of the patterns described in P of EAA — Table Data Gateway, Row Data Gateway, Data Mapper, Unit of Work, Identity Map, Lazy Load, and so on, all catalogue what are now different components of modern-day ORMs.

So when the Repository pattern talks about mediating between the domain and “data mapping layers,” it isn’t referring to your ORM as a whole, as most developers seem to assume, but to just one component of your ORM — specifically, the component that copies data from the results of the generated SQL query into your entities. This mediating layer is also an element of functionality provided by modern ORMs.

For example, Entity Framework’s DbSet<T> is a Repository. So too is NHibernate’s ISession, with methods such as QueryOver<T>().

So what is the wrapper class that people write around their ORMs then, the one that they tend to refer to as a Repository? A more accurate term for this is, in actual fact, a Repository Facade.

It’s important to draw the distinction, especially with the debate around whether this pattern has any value or not. Referring to your ORM itself as a Repository makes it easy for people to make the conceptual leap that allows them to just plug Entity Framework straight into their business service classes without the additional layer of abstraction, but on the other hand it can cause a bit of confusion if you then start saying that “the Repository pattern is harmful.” That’s why I’m now being careful to use the term “Repository” to refer to Entity Framework, NHibernate or the RavenDB client itself, and the term “Repository Facade” to refer to the practice of adding an extra abstraction layer around it.

Posted at 08:00 on 9 October 2014.
Tagged: design, n-tier deconstructed

The key feature of Dolstagis.Web is that it builds the concept of feature switches and Branch by Abstraction right into its core. To anyone who recalls the position that I adopted three years ago in The Great Feature Branch Debate, this will no doubt come as a bit of a surprise. Am I jumping the shark here?

Not really. I may have been objecting at the time to the black-and-white “feature branches bad, feature toggles good” line being adopted by Martin Fowler and his ThoughtWorks colleagues, but my own opinion is somewhat more nuanced than the polar opposite of “feature toggles bad, feature branches good.” There are actually a lot of clever things that you can do with feature toggles:

• You can set them to flip at a specific date and time, for example, if you want to launch something at a conference.
• You can have them go by IP address. For example, country-specific features.
• You can use them for A/B testing.
• You can use them to turn off features and degrade your site’s performance gracefully under load.
• You may occasionally have a feature that needs to be turned off in order to perform a specific upgrade.

The problem with feature switches is that when they are used as an alternative to branching and merging, you are regularly deploying code into your production environment that you know for a fact to be buggy, immature and incorrect. If your feature switches aren’t properly architected, this can end up being exposed to the general public — with disastrous results. Breaking your tasks down into small user stories can help, but this is not always the case.

The commonest mistake here is that your feature switch doesn’t switch everything. On an ASP.NET web application, for example, it may switch out your controllers, or your routes, but you will still have the same views, the same JavaScript, the same CSS and the same images and other static assets available for both the “on” and “off” states. It’s difficult to turn static assets on or off when by default they’re all served up by IIS from your web application’s filespace.

Dolstagis.Web gets round this problem quite simply by requiring you to expose your static files and your views explicitly in your feature definition classes. If it hasn’t been added, it won’t be shown: it’s as simple as that. In fact, you aren’t even limited to using your application filespace: you could just as easily include your static files and views as resources within your assembly. You can even have the feature that you are toggling in a separate assembly altogether, and modify your build process so that your production version doesn’t even have a copy of the immature and buggy code.

Of course this doesn’t guarantee you that you’ll get your feature switches right, and there are still ways in which you can get them wrong, but hopefully this approach will help to make it easier to avoid running into problems.

Posted at 08:00 on 2 October 2014.
Tagged: design, feature branches, feature switches, pet projects