[prev in list] [next in list] [prev in thread] [next in thread]
List: kde-devel
Subject: Re: Nepomuk in 4.13 and beyond
From: Ignacio Serantes <kde () aynoa ! net>
Date: 2013-12-24 11:11:56
Message-ID: CAKbQbApJa45wej511rYdh-RNDhGFPTujt0DcZWLmQQ4z1vtEPQ () mail ! gmail ! com
[Download RAW message or body]
[Attachment #2 (multipart/alternative)]
On Tue, Dec 17, 2013 at 8:54 PM, Kevin Krammer <krammer@kde.org> wrote:
> On Tuesday, 2013-12-17, 20:09:21, Ignacio Serantes wrote:
> > Hi
> >
> > On Tue, Dec 17, 2013 at 6:33 PM, Kevin Krammer <krammer@kde.org> wrote:
> > > Hi Ignacio,
> > >
> > > On Tuesday, 2013-12-17, 17:55:53, Ignacio Serantes wrote:
>
> > > Maybe you could elaborate how a system service would facilitate this
> kind
> > > of
> > > sharing while a session service does not.
> >
> > Well, because I can't do a query if service is not running unless I'm
> > replicating all my data over all my computers. If I have a big database
> > this would be impossible because you must wait for a long time until your
> > data was synchronized. If you are running this as a service you could
> > connect to that server and you don't need to synchronize this data. A big
> > database in a mobile or tablet could be problematic because is common you
> > have gigas in your PCs but megas in your mobile devices.
>
> It seems that you are describing a remote server now, which would be the
> same
> case independent on whether the local service would run in the user
> session or
> as a system service, no?
>
A remote server is the origin of my petitions, I like to install Baloo in a
server, real or virtual because with Nepomuk it's impossible.
>
> > > > Obviously this was my petitions :), I'm not interested at all in
> > > > Akonadi,
> > > > is old computing too for people who works in only one computer, and I
> > >
> > > can't
> > >
> > > > share my Nepomuk's data with all my devices so I think a software
> > >
> > > developed
> > >
> > > > in 2013 supports 2013 software requirements.
> > >
> > > I think you might have some misconceptions about Akonadi but you are
> > > welcome
> > > to proof me wrong :)
> > >
> > > Probably, when Akonady has a check to disable it maybe I change my mind
> > > ;).
>
> Which neatly proofs my point about misconceptions :)
> Akonadi is started on a need basis by its clients. In order to "disable" it
> you simply don't use any of its clients (which would be useless anyway
> without
> data).
>
> I feel a deja vu. Ok, I'm not interested at all in Akonadi and I don't
want waste my time arguing about how good or bad it is anymore. This is
about Baloo :).
> > > A uniform and data type agnostic access layer sounds pretty state of
> the
> > > art
> > > to me.
> >
> > Yes, but a single user one running in your user session and not as a
> > service when you don't want it not.
>
> Not quite understanding your "but". It is exactly running in a user
> session,
> as a service, and not running when you don't want to.
>
> > With so many years of development I'm
> > assuming Akonadi works well for a single user in a single user computer
> > with only one session opened :). But this is not about Akonadi, is about
> > Baloo :).
>
> True, but you brought it up for whatever reason and so I am trying to
> understand that reason. Somehow it must have had some context for you, no?
>
> > > > For me it will be terrific when I tag, comment or rate a file in one
> of
> > >
> > > my
> > >
> > > > devices and automatically this information will be available in all
> my
> > > > devices and if Baloo works at user level this will be impossible.
> > >
> > > I don't see why this would be impossible.
> > > That's like saying "I'd like my email to be marked unread on all my
> > > computer
> > > automatically but if Akonadi wokrs at the user level this will be
> > > impossible".
> > > Which would be demonstratably wrong :)
> >
> > Yes, you are right, but this is achieved with a synchronization method.
> As
> > I commented in my first reply I see problems synchronizing this
> information
> > because different hardware in your devices.
>
> Seems to work quite well for PIM data (emails, contacts, calendars).
>
You can park a plane but this is not a proof than planes are build to be
driven for a highway. I spend my first years synchronizing data before
internet and I have bad memories about this years and all the problems
related with synchronization.
>
> > Let's try an example. I have over 50.000 emails in my Akonadi database
> with
> > the synchronization approach if I want to search for a mail tagged as "My
> > tag" and dated two years ago I will need all 50.000 emails metadata in
> all
> > my devices or search will fail.
>
> Depends on how your search is implemented and how your data is stored.
> If all that email is local then accessing the data should be a problem.
> Might
> be more efficient when indexed but not necessarily mandatory,
> If all that email is on an IMAP server, search could be delegated to it
> when
> the server is reachable and fall back to whatever is currently cached when
> not.
>
> > In my case I have a tablet, two mobiles (Android and iPhone), and four
> > computers (one in my job) so I need all this data synchronized in 7
> devices
> > or this metadata is useless because is not reliable. The worst part is
> when
> > I'm in a customer office working and the only method to connect to the
> > world is a client computer because 3G is not working so, in brief, my
> > salvation is a server where all my mail is stored with my metadata and
> > accessible from any kind device with a browser.
>
> And that is orthogonal to a local service for providing access to the
> user's
> applications how?
>
> As far as I can see it would only be benefitial if only one program would
> cause network traffic for every single data interaction, especially in the
> 3G
> case.
>
>
Sorry. English is not my language and I don't understand you. I'm talking
about a server implementation advantages. If I tagged 10000 mails in my PC
this is done in the server so my mobile device did not waste any byte
downloading that information.
> Cheers,
> Kevin
> --
> Kevin Krammer, KDE developer, xdg-utils developer
> KDE user support, developer mentoring
>
>
> >> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to
> unsubscribe <<
>
>
And I think is time to stop our noise in the mailing list. If you like I
have no problems in arguing about server approach or synchronization
approach using private mail.
--
Best wishes,
Ignacio
[Attachment #5 (text/html)]
<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Tue, Dec 17, 2013 \
at 8:54 PM, Kevin Krammer <span dir="ltr"><<a href="mailto:krammer@kde.org" \
target="_blank">krammer@kde.org</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0px 0px 0px \
0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"><div \
class="im">On Tuesday, 2013-12-17, 20:09:21, Ignacio Serantes wrote:<br>
> Hi<br>
><br>
> On Tue, Dec 17, 2013 at 6:33 PM, Kevin Krammer <<a \
href="mailto:krammer@kde.org">krammer@kde.org</a>> wrote:<br> > > Hi \
Ignacio,<br> > ><br>
> > On Tuesday, 2013-12-17, 17:55:53, Ignacio Serantes wrote:<br>
<br>
</div><div class="im">> > Maybe you could elaborate how a system service would \
facilitate this kind<br> > > of<br>
> > sharing while a session service does not.<br>
><br>
> Well, because I can't do a query if service is not running unless \
I'm<br> > replicating all my data over all my computers. If I have a big \
database<br> > this would be impossible because you must wait for a long time \
until your<br> > data was synchronized. If you are running this as a service you \
could<br> > connect to that server and you don't need to synchronize this \
data. A big<br> > database in a mobile or tablet could be problematic because is \
common you<br> > have gigas in your PCs but megas in your mobile devices.<br>
<br>
</div>It seems that you are describing a remote server now, which would be the \
same<br> case independent on whether the local service would run in the user session \
or<br> as a system service, no?<br></blockquote><div>A remote server is the origin of \
my petitions, I like to install Baloo in a server, real or virtual because with \
Nepomuk it's impossible.</div><div> </div><blockquote class="gmail_quote" \
style="margin:0px 0px 0px \
0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
<div class="im"><br>
> > > Obviously this was my petitions :), I'm not interested at all \
in<br> > > > Akonadi,<br>
> > > is old computing too for people who works in only one computer, and \
I<br> > ><br>
> > can't<br>
> ><br>
> > > share my Nepomuk's data with all my devices so I think a \
software<br> > ><br>
> > developed<br>
> ><br>
> > > in 2013 supports 2013 software requirements.<br>
> ><br>
> > I think you might have some misconceptions about Akonadi but you are<br>
> > welcome<br>
> > to proof me wrong :)<br>
> ><br>
> > Probably, when Akonady has a check to disable it maybe I change my mind<br>
> > ;).<br>
<br>
</div>Which neatly proofs my point about misconceptions :)<br>
Akonadi is started on a need basis by its clients. In order to "disable" \
it<br> you simply don't use any of its clients (which would be useless anyway \
without<br> data).<br>
<div class="im"><br></div></blockquote><div>I feel a deja vu. Ok, I'm not \
interested at all in Akonadi and I don't want waste my time arguing about how \
good or bad it is anymore. This is about Baloo :).<br><br> </div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px \
0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"><div \
class="im"> > > A uniform and data type agnostic access layer sounds pretty \
state of the<br> > > art<br>
> > to me.<br>
><br>
> Yes, but a single user one running in your user session and not as a<br>
> service when you don't want it not.<br>
<br>
</div>Not quite understanding your "but". It is exactly running in a user \
session,<br> as a service, and not running when you don't want to.<br>
<div class="im"><br>
> With so many years of development I'm<br>
> assuming Akonadi works well for a single user in a single user computer<br>
> with only one session opened :). But this is not about Akonadi, is about<br>
> Baloo :).<br>
<br>
</div>True, but you brought it up for whatever reason and so I am trying to<br>
understand that reason. Somehow it must have had some context for you, no?<br>
<div class="im"><br>
> > > For me it will be terrific when I tag, comment or rate a file in one \
of<br> > ><br>
> > my<br>
> ><br>
> > > devices and automatically this information will be available in all \
my<br> > > > devices and if Baloo works at user level this will be \
impossible.<br> > ><br>
> > I don't see why this would be impossible.<br>
> > That's like saying "I'd like my email to be marked unread on \
all my<br> > > computer<br>
> > automatically but if Akonadi wokrs at the user level this will be<br>
> > impossible".<br>
> > Which would be demonstratably wrong :)<br>
><br>
> Yes, you are right, but this is achieved with a synchronization method. As<br>
> I commented in my first reply I see problems synchronizing this information<br>
> because different hardware in your devices.<br>
<br>
</div>Seems to work quite well for PIM data (emails, contacts, \
calendars).<br></blockquote><div> </div><div>You can park a plane but this is not a \
proof than planes are build to be driven for a highway. I spend my first years \
synchronizing data before internet and I have bad memories about this years and all \
the problems related with synchronization.</div>
<div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px \
0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
<div class="im"><br>
> Let's try an example. I have over 50.000 emails in my Akonadi database \
with<br> > the synchronization approach if I want to search for a mail tagged as \
"My<br> > tag" and dated two years ago I will need all 50.000 emails \
metadata in all<br> > my devices or search will fail.<br>
<br>
</div>Depends on how your search is implemented and how your data is stored.<br>
If all that email is local then accessing the data should be a problem. Might<br>
be more efficient when indexed but not necessarily mandatory,<br>
If all that email is on an IMAP server, search could be delegated to it when<br>
the server is reachable and fall back to whatever is currently cached when<br>
not.<br>
<div class="im"><br>
> In my case I have a tablet, two mobiles (Android and iPhone), and four<br>
> computers (one in my job) so I need all this data synchronized in 7 devices<br>
> or this metadata is useless because is not reliable. The worst part is when<br>
> I'm in a customer office working and the only method to connect to the<br>
> world is a client computer because 3G is not working so, in brief, my<br>
> salvation is a server where all my mail is stored with my metadata and<br>
> accessible from any kind device with a browser.<br>
<br>
</div>And that is orthogonal to a local service for providing access to the \
user's<br> applications how?<br>
<br>
As far as I can see it would only be benefitial if only one program would<br>
cause network traffic for every single data interaction, especially in the 3G<br>
case.<br>
<div class=""><div class="h5"><br></div></div></blockquote><div><br></div><div>Sorry. \
English is not my language and I don't understand you. I'm talking about a \
server implementation advantages. If I tagged 10000 mails in my PC this is done in \
the server so my mobile device did not waste any byte downloading that \
information.</div>
<div><br></div><div> <br></div><blockquote class="gmail_quote" style="margin:0px 0px \
0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"><div \
class=""><div class="h5">
Cheers,<br>
Kevin<br>
--<br>
Kevin Krammer, KDE developer, xdg-utils developer<br>
KDE user support, developer mentoring<br>
</div></div><br><br>
>> Visit <a href="http://mail.kde.org/mailman/listinfo/kde-devel#unsub" \
target="_blank">http://mail.kde.org/mailman/listinfo/kde-devel#unsub</a> to \
unsubscribe <<<br> <br></blockquote></div><br>And I think is time to stop our \
noise in the mailing list. If you like I have no problems in arguing about server \
approach or synchronization approach using private mail.<br clear="all"><div><br>
</div>-- <br>Best wishes,<div>Ignacio</div><div><br></div>
</div></div>
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic