On Sun, May 5, 2019 at 12:04 AM Zhuoyun Wei <wzy...@wzyboy.org> wrote:
> On Sat, May 4, 2019, at 15:00, Stefano Zacchiroli wrote: > > > Python data classes might come in handy here: > > https://docs.python.org/3/library/dataclasses.html > > I like dataclasses. It has the same usage interface and repr as > namedtuples, but with the flexibility of classes. There will be no more > awkward `txn = txn._replace(narration='foo')`, just `txn.narration = `foo`. > The flexibility you like from dataclasses is specifically that which I'd want to avoid. The "danger" is to start mutating instances as a side-effect of calling functions. While it's possible to carefully use the mutable interface of dataclasses only on brand new instances, the awkward _replace() interface of namedtuple makes that even clearer. In an ideal world, all mutation should occur upon construction. In other words, unlike most Python developers, I am - utterly uncaring for classes (1) - terrified of side-effects and mutation As such, I'd make all dataclasses frozen... Furthermore, I believe in a kind of "bill of rights" for programmers: Programmers should feel free to use any subset of a language as they see fit. A great new tool would be one that programmatically allows one to enforce such constraints over sections of a codebase (e.g. "under this directory, no classes are allowed"). I wish for something like that in Python and C++ and Go. Using a subset of a language doesn't make it unidiomatic. In particular, avoiding the use of mutation in Python code doesn't make it any less Pythonic. You have to consider this in the context of the evolving nature of programming languages, and the many regrets of their creators. Python has way too many features. The only problem is that it's in Python 3.7. AFAIK some distros are stuck > with Python 3.6... > -- > Zhuoyun Wei > > -- > You received this message because you are subscribed to the Google Groups > "Beancount" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to beancount+unsubscr...@googlegroups.com. > To post to this group, send email to beancount@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/beancount/cb585c68-84d0-4f3c-be43-f2deb65f5493%40www.fastmail.com > . > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Beancount" group. To unsubscribe from this group and stop receiving emails from it, send an email to beancount+unsubscr...@googlegroups.com. To post to this group, send email to beancount@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/CAK21%2BhMArb9BH4AyZTvj8Xc41KXFTgUgwtwpdMj8vFsTB_74kA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
