<\body> Imagine that you want to change the rendering of a given tag, like . As a general rule, provides a set of well-chosen macros which can be customized by the user so as to obtain the desired effect. For instance, as we have seen , you should use modify one of the macros , or in order to customize the rendering of and all other theorem-like environments. However, in some cases, it may not be clear which ``well-chosen'' macro to customize. If we just wanted to change the presentation of lemmas and not of any other theorem-like environments, then we clearly cannot modify , or . In other cases, the user may not want to invest his time in completely understanding the macro hierarchy of , and find out about the existence of , and . So imagine that you want all lemmas to appear in red. One thing \ you can always do is copy the original definition of lemmas in a safe place and redefine the lemma macro on top of the original definition: <\tm-fragment> <\inactive*> > >>>> Alternatively, if only the text inside the lemma should be rendered in red, then you may do: <\tm-fragment> <\inactive*> > >>>> Of course, you have to be careful that the name is not already in use. Another frequent situation is that you only want to modify the rendering of a tag when it is used inside another one. On the web, the language () provides a mechanism for doing this. In , you may simulate this behaviour by redefining macros inside a . For instance, imagine that we want the inter-paragraph space inside lists inside theorem-like environments to vanish. Then we may use: <\tm-fragment> <\inactive*> > |>>>||>>>>>> On the one hand side, this mechanism is a bit more complex than , where it suffices to respecify the attribute of lists inside theorems. On the other hand, it is also more powerful, since the macro applies to all theorem-like environments at once. Furthermore, if the above mechanism is to be used frequently, then real hackers may simplify the notations using further macro magic.