Separate convert and kernel logic One way to significantly cut down on compile-times and testing complexity is to isolate the logic for con- verting to the desired type and the computation. This lets the types just like any other data. Warning Metaprogramming is a powerful tool, but it introduces complexity that can make code more difficult to understand. For example, it can be surprisingly hard to get into Julia code. For manipulating Julia code, use the Expr data structure directly to avoid the complexity of how Julia syntax is parsed. The best uses of metaprogramming often implement most of their

Separate convert and kernel logic One way to significantly cut down on compile-times and testing complexity is to isolate the logic for con- verting to the desired type and the computation. This lets the types just like any other data. Warning Metaprogramming is a powerful tool, but it introduces complexity that can make code more difficult to understand. For example, it can be surprisingly hard to get into Julia code. For manipulating Julia code, use the Expr data structure directly to avoid the complexity of how Julia syntax is parsed. The best uses of metaprogramming often implement most of their

Separate convert and kernel logic One way to significantly cut down on compile-times and testing complexity is to isolate the logic for con- verting to the desired type and the computation. This lets the types just like any other data. Warning Metaprogramming is a powerful tool, but it introduces complexity that can make code more difficult to understand. For example, it can be surprisingly hard to get into Julia code. For manipulating Julia code, use the Expr data structure directly to avoid the complexity of how Julia syntax is parsed. The best uses of metaprogramming often implement most of their

Separate convert and kernel logic One way to significantly cut down on compile-times and testing complexity is to isolate the logic for con- verting to the desired type and the computation. This lets the types just like any other data. Warning Metaprogramming is a powerful tool, but it introduces complexity that can make code more difficult to understand. For example, it can be surprisingly hard to get into Julia code. For manipulating Julia code, use the Expr data structure directly to avoid the complexity of how Julia syntax is parsed. The best uses of metaprogramming often implement most of their

Separate convert and kernel logic One way to significantly cut down on compile-times and testing complexity is to isolate the logic for con- verting to the desired type and the computation. This lets the types just like any other data. Warning Metaprogramming is a powerful tool, but it introduces complexity that can make code more difficult to understand. For example, it can be surprisingly hard to get into Julia code. For manipulating Julia code, use the Expr data structure directly to avoid the complexity of how Julia syntax is parsed. The best uses of metaprogramming often implement most of their

Separate convert and kernel logic One way to significantly cut down on compile-times and testing complexity is to isolate the logic for con- verting to the desired type and the computation. This lets the types just like any other data. Warning Metaprogramming is a powerful tool, but it introduces complexity that can make code more difficult to understand. For example, it can be surprisingly hard to get into Julia code. For manipulating Julia code, use the Expr data structure directly to avoid the complexity of how Julia syntax is parsed. The best uses of metaprogramming often implement most of their

Separate convert and kernel logic One way to significantly cut down on compile-times and testing complexity is to isolate the logic for con- verting to the desired type and the computation. This lets the types just like any other data. Warning Metaprogramming is a powerful tool, but it introduces complexity that can make code more difficult to understand. For example, it can be surprisingly hard to get into Julia code. For manipulating Julia code, use the Expr data structure directly to avoid the complexity of how Julia syntax is parsed. The best uses of metaprogramming often implement most of their

Separate convert and kernel logic One way to significantly cut down on compile-times and testing complexity is to isolate the logic for con- verting to the desired type and the computation. This lets the types just like any other data. Warning Metaprogramming is a powerful tool, but it introduces complexity that can make code more difficult to understand. For example, it can be surprisingly hard to get into Julia code. For manipulating Julia code, use the Expr data structure directly to avoid the complexity of how Julia syntax is parsed. The best uses of metaprogramming often implement most of their

Separate convert and kernel logic One way to significantly cut down on compile-times and testing complexity is to isolate the logic for con- verting to the desired type and the computation. This lets the types just like any other data. Warning Metaprogramming is a powerful tool, but it introduces complexity that can make code more difficult to understand. For example, it can be surprisingly hard to get into Julia code. For manipulating Julia code, use the Expr data structure directly to avoid the complexity of how Julia syntax is parsed. The best uses of metaprogramming often implement most of their

Separate convert and kernel logic One way to significantly cut down on compile-times and testing complexity is to isolate the logic for con- verting to the desired type and the computation. This lets the types just like any other data. Warning Metaprogramming is a powerful tool, but it introduces complexity that can make code more difficult to understand. For example, it can be surprisingly hard to get into Julia code. For manipulating Julia code, use the Expr data structure directly to avoid the complexity of how Julia syntax is parsed. The best uses of metaprogramming often implement most of their