Tanglu B-Release name voting will happen soon
inkapnea at gmail.com
Sun Feb 2 17:38:37 EST 2014
sorry for the late reply, it was left unfinished in Draft :-(
On Tue, Jan 21, 2014 at 7:46 PM, <ca2013 at arcor.de> wrote:
>> Basically, the common user is not concerned with the release date.
>> Does somebody care about the year/month Windows 7 or 8 were released?
>> I think no, he wants to know only 2 things:
>> - if he's running the last, the second last or an obsolete version.
>> - if the version installed on his machine is still supported.
> To know this, you have to know, or look up, when version X was released.
> And this differs for every distro and different release/support schedules.
> Whereas, when you first sit down at machine with a version 2014-02, that
> says all you need to know that it is running a current version that is less
> then a year old, or not.
Indeed, you prove only that it was released less than a year ago (this
is accurate), everything else is relative and the user can't avoid
doing the homework. (Proposal for Tanglu: "release date" and
"security updates till..." could be supplied in the About box).
Here goes another example: one could easily assume that Netrunner
13.12 has never packages than Kubuntu 13.10 because it was released
two months later. Slightly updated packages on the ISO yes, but after
the first update you get the same versions of the kernel, KDE, office
package... To add some confusion, the first is supported for 9 months,
the later 7.
A different example: Kubuntu 12.10 was released in October 2012 while
Debian 7 in May 2013 (so I'll call it Debian 13.05). From the version
it's obvious it has been released seven months after Kubuntu, one
would expect to find newer (or same) versions of the software on
Let's compare some regularly evolving components:
Kubuntu 12.10: Debian 13.05:
kernel 3.5 kernel 3.2
KDE 4.9 KDE 4.8
LibreO 3.6 LibreO 3.5
Even standardizing the naming across distributions to year-month would
bring only a minimal advantage to a technical user when comparing
different distributions because every project has it's own goals,
features and different development rates.
For a common user every distribution is seen as an entity that is
progressively pursuing it's vision, from a release to a release. This
is why progressive numbers give the idea of a product that is evolving
to the better.
>> Windows 8.1 is clearly a minor update to 8 as it's version number suggests.
>> Sorry for using the ?bad guy? as an example but it fits the error
>> prone common user paradigm :-)
> No problem. It may make a difference that they don't have a release twice
> a year, though.
Sure :-) but examples can't be 100% accurate, if I mentioned Mint or
Fedora it could still be possible to find flaws with the reasoning
behind. I mentioned Windows because it's closer to another topic that
I wanted to address - branding.
>> When I'll hear the news that Tanglu 5 is out, I'll check my current
>> install, do the math and say: "I have still the version 3, so I'm two
>> versions behind, it's time to upgrade!"
>> Drawing conclusions from numbers 1.2, 1.8, 2.2 and 2.8 is not so
> Hm well, when one reads an annoucement about the Tanglu release 2014-02,
> nobody has to know or check anything to know it is new (at the
> moment) and old (next year).
As a close date gap date could give a POSITIVE impression about a
fresh product, a far gap could give a NEGATIVE impression about an
outdated and no longer relevant product (despite it could be an LTS).
Those are things to consider when having to deal with general users.
People generally just like to label products and product versions with
good or bad - "don't install v3, it's worse than v2, just wait for
v4". Naming with dates could have a negative side effect because it's
not so easy to label them that way and Tanglu could create apathy to
> If one checks the version that is currently installed (imagine 2013-02),
> one knows how old it is without having to know the particulare release schedule
> of the distro, the date of the first release and to do the math.
My point is that a typical enthusiast user wants to know if he missed
something, clear progressive numbers grow in him a sense of
confidence, it's important to not overlook this aspect.
This discussion underestimated the power of branding - as names,
images and colors, also numbers are part of a brand. Don't forget that
Tanglu's target users are regular users and enthusiasts. I'm not
saying that those are dumb but that a simple number (as a short simple
name) is more incisive, is easier to remember and pronounce when
referring to a product.
Can you imagine a conversation between two guys at the pub comparing
different versions of Tanglu, their ups and downs, constantly
mentioning things like 2014-02, 2014-08, 2015-02? :-)
>> The Ubuntu numbering convention is good and indicative and we have
>> become accustomed over the years. While having more distributions
>> named this way (YY.MM and YYYY.MM) could allow direct comparisons with
>> releases of other distros, the drawback could become bigger for common
>> users when looking for help on the web. If you have different patterns
>> to remember: Ubuntu (YY.04, YY.10), Tanglu (YY.02, YY.08), some eight
>> month cycle distro (YY.01, YY.09, YY.05), some random-cycle distro
>> (YY.09, YY.12) how are you going to remember all those numbers
>> correctly without the risk of messing them? "I need help with my
>> install of Tanglu 14.04." "What? Are you sure?"
>> Also, this is not very useful when comparing Debian and Tanglu versions.
> The ubuntu numbering isn't that clear at all (as you show with your examples),
The ubuntu numbering is good, not great.
When Ubuntu was announced, it was clearly stated that it will have
releases every 6 months (if I remember correctly, at that time not a
single distribution had a regular time based release, or at least not
twice a year). Regular 6 months releases were a BIG news, so to
emphasize this uniqueness and stand out of the crowd Canonical adopted
this new numbering scheme. At this time it was a clever marketing
10 years later I don't really see much benefit of adopting this or the
modified 2014-02 pattern, except for some expert distribution.
The Ubuntu scheme is OK to stay there for Ubuntu and spins. It has
it's pros and we have been accustomed to it over the years.
>> Tanglu is not a rolling distribution. Year and month numbering schemes
>> seem more suitable for rolling distributions where install ISOs are
>> periodical snapshots of the archive state and are rarely referred when
>> asking for support (except install problems).
> I don't really understand. Don't you see how it is also clearer to refer to Tanglu ISOs by
> date instead of version X. (Say if a user finds a disk in his drawer.)
The date on a rolling distribution ISO is important only to give you a
rough estimate about the number of packages that you'll have to
download on the first update (and eventual number of manual changes
needed). After that operation your OS will be up to date, no need
anymore to mention your install ISO version or care when you will stop
getting updates. There are no branches, every installation becomes
synchronized with The Repo, and your installation is supported
I think that naming the first release with 1 is the least "bug prone"
scheme for the Tanglu case (considering target users and possible
shifts in release dates).
Relatively Off-topic: in the meantime at the openSUSE camp a
discussion is ongoing about a temporary(?) shift from the 8 months to
a yearly release cycle. Their unique numbering scheme was finely
synched with the 8 month assumption (it made sense to me despite not
being self descriptive). This will surely have an impact to the
numbering scheme users were used to.
This is one more reason why I think Tanglu should adopt the simplest
scheme and avoid possible future exceptions to the numbering rules.
More information about the Tanglu-devel