On Tue, 16 Jul 2013 09:48:06 +0000, Carlos E. R. wrote:
> On 2013-07-16 05:46, vazhavandan wrote:
>>
>> don’t know as to how nntp works but I am curious as to why nntp doesn’t
>> send updates of posts as mails to users
>
> That is precisely what I ask for and they are saying “NO”, and they
> laugh and ridicule us instead. :-/
Well, no, but at least for my part, I’ve tried to explain why some things
are more difficult to change than it might seem from the ‘outside’.
It is wearing to hear about minor little things (and let’s face it, 3
posts out of 7,000 is a minor thing) as if they’re show-stopping
problems.
All of the staff here (besides Kim) volunteer our time. This is what we
do in our spare time. We’re not paid or compensated in any way for what
we do. So these kind of nits, while raising them is fine, insisting that
they’re serious problems that must be addressed right now or the world is
going to end (and with that kind of hyperbole) really pushes buttons.
And there are a few people who use these forums who seem to delight in
finding ways to push those buttons, and then get mad when we try to
explain anything.
For me personally, it’s to the point where I’m tempted to just reply to
every such report with nothing but “thanks for the report” or “your
concerns are noted”, but because of how I’m wired, I have a hard time
limiting my replies like this.
> Not via email, but via nntp, as another post with the modified post.
Again, from the outside and not knowing how the gateway actually works,
the complexity of the database behind the web side, and everything, it
does look like a simple problem to solve. But because of how the
gateway was written originally (not by us) and the original maintainer no
longer maintains it - and in fact we’re one of two or three
implementations of vBulletin worldwide that even uses it any more -
because the upgrade to vBulletin v4 it completely broke the gateway, but
I managed to figure out how to get it working again.
The vBulletin database schema is very complex. So I hesitate to do
anything that could impact the integrity of that database. Fixing this
minor issue could result in unexpected database corruption, and you can
bet that people would get very angry if the forums became unreliable or
unusuable because the database was corrupted. Yes, I test on a sandbox
installation that I have myself (I gate to an INN server on my home
systems), but I can’t do the scale testing needed to deal with the kind
of message flow that we use.
The same gateway is used on the SUSE Forums, Novell Forums, and NetIQ
forums as well - so changes that I make to the gateway potentially affect
those as well - which would affect paying customers, too.
So messing with it to fix such a minor problem is about managing risk -
and in my judgement, the changes needed have an unacceptable risk to the
message database itself without a fair bit of work and testing. It isn’t
just about writing a fix - it’s about testing it enough, and regressing
the change to ensure we don’t completely screw the forums.
Now, if it seems that we get a little touchy when a report comes in like
this that really is not a huge issue (forum help is “best effort” for
any of us answering questions - and ultimately the risk falls to those
asking the questions, not those providing the answers), maybe you’ll
understand a little better why that is.
Jon’s response was maybe a bit OTT, but it’s indicative of the type of
frustration that these kinds of “demands” that we fix minor issues put on
all of the staff.
There are still more important things to be looking at, like why NNTP
posts don’t trigger e-mail alerts for subscriptions consistently.
But ultimately, and I know you and DD both get tired of hearing this, NNTP
is a secondary interface. It’s provided for your convenience. It isn’t
perfect, and we have repeatedly stated that. It ain’t gonna be perfect,
and if you need perfection, you’re going to have to use the web interface
and live with its limitations.
Every once in a while, we have to have a debate about whether or not it’s
worth keeping the NNTP interface, with its warts and all. In spite of
its shortcomings, I always want to see it maintained, which means that
when I do the risk analysis for whether or not an issue should be looked
into or fixed, I’m looking at it with an eye towards “if this screws
things up, would it be the nail in the NNTP coffin?”, and if I judge the
risk is at an unacceptable level, then I’m not going to look at making a
change. I would far rather have an imperfect NNTP gateway solution than
have one that breaks things to the point that we end up having to restore
from a backup and decide not to support the interface any more at all.
So yes, every time you start pushing the buttons for fixing something
that’s not got a significant impact, it’s probably not going to get
addressed, and if you keep pushing, you’re going to frustrate members of
staff here. We’re only human, after all. And when someone goes to the
lengths that DD did in his hypothetical “critical piece of information
wasn’t in the original post that went through the gateway and the advice
given blew their system away because we were missing that critical piece
of information that would have saved their system” post, well, that
really don’t make the situation any better. We can all construct wild
hypothetical situations where a system has been completely destroyed
because someone forgot something. I can also construct hypothetical
situations where the person reading the answer didn’t understand it
correctly, or the person answering didn’t explain something clearly
enough, or, or, or, or - and those hypothetical situations don’t do us
one bit of good. They just end up being “talking a lot without saying
anything”. (Cue the counter now the date I’m accused of saying that
about some specific forum member, another frequent SOP I see from certain
quarters - which I should also start a counter for being accused of
applying to someone specific - neither of which I’m doing, BTW - I’m
speaking in generalities.)
So now I’m going to go and reset the “days since having to explain in
great and excruciating detail why some minor issue isn’t going to get
addressed” counter to 0 again. 
I actually do know what I’m doing with the gateway, and I do actually
care what the community sees as problems, and as much as possible I want
to see issues resolved (we all want continual improvement). But
mountains out of molehills doesn’t move things forward. Now, I can waste
more time explaining stuff in excruciating detail, or I can go and get
some paid work done so that I might be able to afford some free time to
look into problems that affect more than 3 posts out of 7,000.
Jim
Jim Henderson
openSUSE Forums Administrator
Forum Use Terms & Conditions at http://tinyurl.com/openSUSE-T-C
