There's a long history of health controversies surrounding artificial sweeteners, with critics accusing them of causing everything from cancer to diabetes to attention deficit disorder. Aaron Carroll reviews the evidence and shows, pretty clearly, that none of the claims are sound. Bottom line: there's no evidence that artificial sweeteners have adverse health effects, and no serious studies cast doubt on the zero-calorie proposition: consuming zero-calorie artificial sweeteners doesn't cause you to gain weight.

So why are so many people so against artificial sweeteners? I'd like to posit a hypothesis: people keep looking for reasons why artificial sweeteners are bad for you because they dislike the way artificial sweeteners taste. Personally, I think they taste awful. Bitter and harsh, and not really all that sweet. Artificial sweeteners taste a bit different to different people, depending largely on genetics. But we all generally do agree that sugar tastes good. Maybe all of this scare is just people looking for a way to justify their unhealthy preference for sugar over artificial sweeteners.

Ok, here's a question: how do programmers make non-web applications in .Net anymore?

I ask because Microsoft has killed off SQL Server Compact Edition, which I had previously thought was the standard way to make a database for a non-web-based program in .Net. But there's no SQLServerCE assembly in Visual Studio 2013 or 2015, not even as a reference you can add--if you want it, you have to download and install the .dll file manually. Every non-trivial program necessarily involves storing and manipulating data, and I have a hard time believing that most non-web programs don't have relational, query-able data that calls for a database, so what exactly are these non-web programs using?

Ok, obviously Microsoft left a way to have a local non-web relational database--in fact, they've now extended SQL Server Express so that you can create local SQL Server databases that work offline for your applications, and this is the standard (only) way to add a local database in 2013-5 editions of Visual Studio. And at first I was quite pleased with this because you can use the same SQL Server client namespaces, such as System.Data.SqlClient for a local database as you would use for a database on the web server, making it useful for development of web applications.

In other words, local databases in .Net programs are now "real" SQL Server databases rather than an alternative made specifically for offline applications. There are advantages to this--for example, Compact Edition databases did not support stored procedures whereas real databases do. On the other hand, I'm not sure why developers of offline apps would care--stored procedures are really a way of handling data complexity, which is important on a server-side system with data streaming in from disparate sources, but unimportant in an offline app where you have full control over the format of the incoming data. And the difference is not benign: whereas Compact Edition databases were self-contained within the offline apps that used them, these new full databases have a hard dependency on SQL Server Express, an entire separate program which must be installed on every computer before they can run the offline app.

And that's what has me puzzled. I've installed a couple offline programs with local databases onto other peoples' computers, and in every case, on both new Windows 8 operating systems and older ones, it has prompted me to install SQL Server Express, thus revealing two things: 1)that this isn't just something that Microsoft has built into new operating systems, so that the dependency doesn't matter, and 2)no program developers are using Microsoft's new system for local databases. For comparison, this is now the code to create a database:

string filename = @"C:\Users\User\Documents\testdb.mdf";
if (!System.IO.File.Exists(filename))
{
    string databaseName = System.IO.Path.GetFileNameWithoutExtension(filename);
    using (var con = new SqlConnection("Data Source=.\\sqlexpress;Initial Catalog=master; Integrated Security=true;User Instance=True;"))
    {
        if (con.State != ConnectionState.Open) con.Open();
        using (var command = con.CreateCommand())
        {
            command.CommandText =String.Format("CREATE DATABASE {0} ON PRIMARY (NAME={0}, FILENAME='{1}')", databaseName, filename);
            command.ExecuteNonQuery();
            command.CommandText =String.Format("EXEC sp_detach_db '{0}', 'true'", databaseName);
            command.ExecuteNonQuery();
        }
        con.Close();
    }
}

That's a lot harder than before. You now have to connect to the computer's SQL Server instance, create a database and then detach that database to it's own file. Then you can connect to the new database.

But the bigger problem is this: SQL Server Express, as best I can tell, does not support transparent encryption. That's a big problem when you think about it. Sure, you can encrypt data before putting it into the database, but then the DBMS can't tell what it's value is when you query it, which severely complicates a lot of the functionality you'd like a program to be able to have (the only way to do "in the past year" type searches is with a web of surrogate keys; if encryption is done correctly, the only way to find a search term is to query and decrypt all the data one-by-one until it's found). Between the dependency on SQL Server and the lack of usable encryption, there's a strong temptation to continue using the old Compact Edition databases. But that's problematic: if Microsoft is no longer supporting Compact Edition, that means they aren't upgrading encryption algorithms and fixing security bugs.

In conclusion, it is now basically impossible to make a standards compliant offline application using only supported Microsoft frameworks. I get and agree with the web-first paradigm, but come on, that's ridiculous.

Any analysis of political ideologies is going to produce some version of this chart:

Basically, ideologies can be measured in terms of what they define as freedom: Republicans and conservatives generally define freedom in terms of economic freedom, and don't have a problem with regulating what people can do outside the marketplace. Democrats and liberals define freedom in terms of the freedom to do whatever you want outside the marketplace, but don't have a problem regulating how people must conduct business inside the marketplace. Libertarians are defined by the fact they value both equally: freedom includes both the ability to do what you want outside the marketplace and inside it. The fourth box is labeled "hardhats," a term that comes from Paul Krugman, but there's really no one in that box because the only way to get there is to not value any kind of freedom at all--not a popular stance.

Andrew Kaczynski points us to this quote from Rand Paul that makes his ideology quite clear:

"if we tax you at 100% then you’ve got zero percent liberty. If we tax you at 50% you are half slave, half free."

Thus, Rand Paul places zero weight on non-economic freedoms, and highly values economic freedoms. That puts him squarely in the Republican/conservative box.

The absolutism of Rand Paul's conservatism is quite remarkable. Consider the hypothetical scenario where the government taxes 100 percent of income but provides free food and other consumption for everyone (nevermind how they pay for it, for the moment). Obviously, there'd be no point to working, and you'd have no freedom to increase your consumption beyond what the government hands out. There's no doubt that the high tax rate would reduce your freedom. But you certainly could enjoy quite a bit of other kinds of freedom in this world--just to take a timely example, you could marry whoever you want, of either gender. Rand Paul's position, then, is that the freedom to marry is insignificant compared to the ability to earn untaxed income. That's pretty solid Republicanism, which has no problem prescribing who people can marry and when they can get divorced, so long as taxes are low.

Incidentally, on Twitter I recently started a debate on a related question and Adam Omizek said this:

@hyperplanes @t0nyyates I think actual slaves might argue they lacked a bit more than a paycheck

— Modeled Behavior (@ModeledBehavior) July 2, 2015

And he's right: the lack of a paycheck is not what defines slavery. To put it differently, merely giving a slave a paycheck does not set him free--in fact, many slave owners did give them paychecks. Clearly, an integral part of the definition of slavery is it's involuntary nature--the compulsory labor and the lack of agency over one's own life. But according to Rand Paul, the inability to work for pay is the same as being forced to work for no pay.