What next? [Was: Re: LiveCD or No LiveCD?]
Alexander E. Patrakov
patrakov at gmail.com
Tue Feb 26 08:08:36 PST 2008
Jeremy Huntwork wrote:
> 20 people expressed their appreciation for the CD, more than half voting
> to keep the project around.
Please subtract the number that want to use the LiveCD to cover LFS bugs, don't
realize the inherent incompatibility of LFS with 64-bit hosts (IMHO, the fact
that LFS doesn't mention it counts as a bug), or don't know how to apt-get
> Also, several either offered to contribute
> or suggested ways in which the project may be improved.
> 2 people explicitly voted to drop the project.
Count me too, unless "transform to something else than a binary distro" doesn't
count as dropping.
All of the below is valid only if the LiveCD subproject is not dropped.
> * Go back to the drawing board, so to speak. Start a new CD from scratch
> that is minimal (and minimal means minimal, not just 'without X') and
> re-define core concepts that the CD will adhere closely to. (For
> example, as proof of the soundness of LFS, the CD should strictly adhere
> to LFS. If we adopt this one aspect, we should also be able to make use
> of ALFS development to produce the CD, instead of maintaining a full set
> of separate scripts.)
Then we need a procedure to deal with feature requests ("reject outright" also
counts as a procedure). Also we need a procedure to determine what counts as a
feature request and what doesn't (e.g., what to do with net-firmware,
scsi-firmware and the numerous kernel patches and extra drivers for new hardware?)
> * As has been suggested from a long time ago, make use of package
> management in the build process, especially for BLFS packages. This
> would allow at least two benefits: an easier development process, and
> greater extensibility/customization.
This is incompatible with the "strictly adhere to LFS" goal, because LFS has no
package management except "rebuild everything once a day". Note that I make no
statement about the relative merit of these two incompatible goals.
> * Add an LFS-style document to the project that teaches how to create a
> LiveCD from scratch.
IMHO, this would line up with the rest of the project nicely, and doesn't
exclude actually providing some binary CD.
> * Devise methods for users to more easily provide feedback and make it
> easier to contribute as a whole.
This also should include stating what kind of feedback is needed. So far, there
were only "please add this package" requests, borderline cases like "old laptops
with the OnTrack disk manager doen't see the partitions with libata" (solved by
adding kpartx) and a few bug reports.
> What are your thoughts on the above? And are there any other
> suggestions, either new ones or ones that I missed?
Alexander E. Patrakov
More information about the lfs-dev