[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

"From Embedian to a Pure Blend - a Bbliography" - does one already exist in some form? - WAS [Re: Towards a Debian for minimalists]



On 7/18/2016 3:15 AM, Petter Reinholdtsen wrote:
...

I am sad you feel that way, and I get the impression that you assume all
recommendations in Debian are correct and appropriate, thought I know
you enough to realize that this is not the case.  In my experience they
are not.  I assume the following knowledge is well known to you, Jonas,
but it might give some relevant context to Richard.

Many years ago, package recommends in Debian were treated as if they
were suggests, ie not installed by default by apt.  This was against the
documented policy, and made recommends mostly useless, and in general
considered a bug in Debian.  Then finally, some years ago, apt was
changed to install recommends by default, and the disk footprint of a
Debian installation exploded.  This was partly because of bogus
recommends (that should have been suggests) and partly because of wanted
recommends that were finally installed as they should be.

It is hard to decide if a recommends fall into the class of bogus or
wanted, and it is up to the package maintainer to decide when in doubt,
but in my view it is some times useful to question if a package listed
as recommended belong in suggests instead.  So if you come across a
package recommending something you suspect should be suggests instead,
do not be afraid of reporting it as a wishlist bug report.
[snip]

I'm taken to task when proposing to disable be default "recommends" for what would be otherwise be considered a routine Pure Blend.

Embedian initially was very minimalist due limited hardware. As hardware constraints were loosened the project was merged with Pure Blends as better utilizing scarce human resources. My personal opinion is mineralization is valuable enough in and of itself for me to invest my time and effort.

Why the bibliography? Someone once said approximately "Those who do not know history are doomed to repeat it." If a bibliography does not exist, I'll have to start piecing one together.

How will Debian benefit?
Along the way I will have to create a structure an what an executable provides that justifies it being in "recommends" of a higher package. and automating it if possible. If I have read https://bugs.debian.org/831502 correctly, what I'm aiming at might have prevented it or at least eased its resolution.

My test case will be a metapackage based on Mate that does a GUI "my way" (apologies to fast food slogan writers).



Reply to: