Re: 48 and 72 ET

[Hans Åberg]

> On 10 Feb 2017, at 10:22, David Kastrup <address@hidden> wrote:
> 
> Hans Åberg <address@hidden> writes:
> 
>>> On 10 Feb 2017, at 00:35, David Kastrup <address@hidden> wrote:
>>> 
>>> Hans Åberg <address@hidden> writes:
>>> 
>>>>> On 10 Feb 2017, at 00:16, David Kastrup <address@hidden> wrote:
>>>>> 
>>>>> So is there any reason people don't use convert-ly when upgrading to
>>>>> a newer version?
>>>> 
>>>> For libraries, you would want to keep track of the changes, but
>>>> running convert-ly and do a diff is a good suggestion.
>>> 
>>> convert-ly -d inserts an updated \version header.
>> 
>> You wouldn't want to do that with library headers, in case something
>> goes wrong. Better to create a separate file, and check file dates, as
>> in a Makefile.
>> 
>>>> Though doing it by hand was quicker, as I remembered the issue and
>>>> which files needed to be fixed.
>>> 
>>> How does _this_ "keep track of the changes"?  Maybe you consider your
>>> biological memory part of the data on your computer?
>> 
>> I cannot parse this.
> 
> You said "you would want to keep track of the changes" as a reason to do
> changes manually rather than by convert-ly.  I tried reading some sense
> into that statement but apparently we have very little common ground
> regarding what we consider making sense.

See the letter from Urs Liska, how he works. It is the human brain that makes 
sure things are correct. Also see the letter by Graham Breed: it is great that 
convert-ly can do the job. But also see the letter by Simon Albrecht: 
convert-ly does not always do it right. So it is good to make that sure, 
especially in a library. Even though it could do it in this case, as you found 
out, I did not need it, and in the past, a very long time ago, I think that 
convert-ly failed, and I haven't needed it since.