Members present: Humbedooh, westei, nandana, tkurz, sschaffert, jfrank, Wikier

----------------
Meeting summary:
----------------

1. Preface

2. triple store

3. Marmotta Website

4. logo
  e. https://issues.apache.org/jira/browse/MARMOTTA-5 (sschaffert, 4)

5. administration interface

6. w3c
  w. https://dvcs.w3.org/hg/ldpwg/raw-file/default/ldp-ucr.html (Wikier, 6)

7. aob


--------
Actions:
--------
look how Apache Maven generates its site from artifacts (sschaffert, 09:31:26)
contact to trademarks@ regarding the logo proposal (Wikier, 09:42:42)
decision on different sizes and formats for the logo (sschaffert, 09:42:57)
decision on other uses of the brand in the software and on the website 
(sschaffert, 09:43:27)
attach screenshot of new admin interface the JIRA issue (tkurz, 09:48:57)
create a jira to track the communication with the ldp wg (Wikier, 09:57:42)
start a discussion at the mailing list about ldp, specially ldp-c (Wikier, 
09:58:59)
continue discussion on LDP issues and contributions from Marmotta (sschaffert, 
10:03:43)

IRC log follows:


# 1. Preface #
09:03:33 [Wikier]: so, what topics do we have?


# 2. triple store #
09:03:50 [Wikier]: project branding, lmf 2.4 release tasks and others
09:04:11 [sschaffert]: yes, and the work on the triple store I am doing right 
now (MARMOTTA-8)
09:04:12 [Wikier]: ok, triple store first
09:04:41 [sschaffert]: so just to summarize again for the newcomers and the 
meeting minutes
09:05:11 [sschaffert]: I am currently in the process of separating out the 
triple store used by the LMF, which was actually developed already in 2009 for 
KiWi
09:05:18 [sschaffert]: so that
09:05:26 [sschaffert]: (1) it can be used independently as a Sesame backend
09:05:41 [sschaffert]: (2) Marmotta can be independent of the triple store and 
use any Sesame backend
09:05:51 [sschaffert]: (including Jena)
09:06:05 [sschaffert]: the KiWi triple store consists of several components
09:06:33 [sschaffert]: - triplestore core: this component has already been 
rewritten, uses direct JDBC access to MySQL, PostgreSQL, H2
09:07:03 [sschaffert]: - triplestore transactions: this covers transaction 
logging, i.e. takes a log of all changes done to the triple store in a 
transaction
09:07:11 [sschaffert]: finished yesterday
09:07:33 [sschaffert]: - triplestore versioning: this I mostly implemented 
yesterday, more detailed report below
09:07:57 [sschaffert]: - triplestore reasoning: a rule-based reasoner (sKWRL) 
operating directly on the database, task for next week
09:08:34 [sschaffert]: - triplestore native SPARQL: SPARQL-to-SQL translation, 
needs to be rewritten, task for last week of January
09:08:48 [sschaffert]: so regarding the versioning component
09:09:04 [sschaffert]: I have finished the implementation more or less, but it 
is completely untested
09:09:26 [sschaffert]: the implementation at the moment allows accessing the 
triple store as a repository at any given time in history
09:10:11 [sschaffert]: so you can say "I want a connection to the state of the 
triplestore on 17th January, 17:00" and you get a read-only connection where 
you can use SPARQL or any other Sesame API tools
09:10:26 [tkurz]: cool feature
09:10:41 [sschaffert]: the background here is that I'd like to integrate into 
the LDP services
09:10:57 [sschaffert]: - certified snapshots, so you can give a link to a 
version that is known to contain correct data
09:11:06 [tkurz]: we can offer it via memento protocol
09:11:28 [sschaffert]: - support for the memento protocol 
#linkhttp://mementoweb.org/
09:11:34 [sschaffert]: - support for the memento protocol #link 
http://mementoweb.org/
09:12:03 [sschaffert]: this is maybe also a contribution to the W3C linked data 
platform WG that we should keep in mind
09:12:18 [sschaffert]: I think as soon as you talk about updates you need to 
take this into account
09:12:33 [sschaffert]: so back to the implementation, this is implemented but 
untested
09:12:41 [sschaffert]: also implemented is listing of versions
09:12:56 [sschaffert]: not implemented is reverting versions (not sure if we 
need this) and removing old versions
09:13:11 [Wikier]: so we have 10 days to prepare a proper test suite
09:13:14 [sschaffert]: task for today is to implement the tests and make sure 
that it works
09:13:33 [sschaffert]: 10 days until the WG meeting at W3C?
09:13:43 [Wikier]: no, for the planned date
09:13:50 [Wikier]: I mean
09:13:51 [sschaffert]: ah, no, planned date is today :)
09:13:56 [sschaffert]: MARMOTTA-8
09:13:56 [Wikier]: ok :-)
09:14:04 [Wikier]: good
09:14:19 [sschaffert]: so my task today is to write the tests and make sure it 
works for all three supported databases
09:14:33 [sschaffert]: there will be two tests:
09:14:41 [sschaffert]: (1) testing if the versions are created properly
09:14:48 [sschaffert]: (2) testing if the snapshot functionality works
09:14:58 [sschaffert]: and this is it regarding the versioning component :)
09:15:27 [tkurz]: okay. when do you plan to merge this into the main branch
09:15:48 [sschaffert]: when the reasoning and SPARQL are finished as well and 
the existing implementation can be replaced
09:15:58 [sschaffert]: this is a huge change because we will drop Hibernate
09:16:11 [sschaffert]: the new implementation directly uses JDBC
09:16:35 [sschaffert]: this has a bit of effect on other LMF modules that also 
use Hibernate, we need to discuss with Jakob
09:16:44 [Wikier]: we should try to get some figure about the gain in 
performance
09:16:48 [sschaffert]: but these modules are probably not part of Marmotta
09:16:56 [Wikier]: yes, user and so on
09:17:18 [sschaffert]: using Hibernate is not precluded, just the triple store 
doesn't need it
09:17:48 [sschaffert]: regarding performance: yes, I am going to do some tests
09:18:03 [Wikier]: cool
09:18:04 [sschaffert]: but real tests I can only do comparing the old complete 
LMF and the new complete Marmotta
09:18:18 [sschaffert]: because in the old version the triple store cannot be 
separated
09:18:26 [Wikier]: well, just to know our ideas really apply
09:18:36 [sschaffert]: but I expect a considerable performance gain
09:18:43 [Wikier]: and it is the right way
09:18:48 [Wikier]: good
09:18:56 [sschaffert]: considerable = factor 4 at least
09:19:19 [Wikier]: really?
09:19:26 [sschaffert]: and I also expect a reduced number of bugs
09:19:26 [sschaffert]: yes
09:19:33 [Wikier]: exactly
09:19:33 [sschaffert]: my guess is based on the following:
09:19:56 [Wikier]: code maintainability is for me enough justification of this 
effort
09:19:56 [sschaffert]: - Hibernate typically produces excessive number of 
queries that are not needed (at least 4 times more queries)
09:20:26 [sschaffert]: - Hibernate typically loads all fields from the tables 
even if they are not needed, so more network I/O
09:20:49 [sschaffert]: so I expect factor 4 at least :)
09:20:57 [Wikier]: let's see...
09:21:03 [Wikier]: welcome, jfrank
09:21:26 [Wikier]: anything else regarding the triple store?
09:21:26 [sschaffert]: any more questions regarding the triple store?
09:21:33 [Wikier]: not from my side
09:22:14 [tkurz]: next topic?
09:22:18 [sschaffert]: so let's move to the next topic


# 3. Marmotta Website #
09:23:11 [Wikier]: ok
09:23:18 [Wikier]: the first draft i available at the svn
09:23:41 [Wikier]: awaiting for infra to register the location at the apache cms
09:24:13 [Wikier]: further details at MARMOTTA-12 and INFRA-5624
09:24:19 [sschaffert]: any way to see it before? I tried with the Perl scripts 
but they did not run out of the box, so I left it
09:24:34 [sschaffert]: (MacOS Perl)
09:24:48 [Wikier]: I don't know in OSX, but it Debian I needed to install some 
extra packages
09:24:57 [Wikier]: perl sucks, yes
09:25:04 [Wikier]: anyway
09:25:12 [sschaffert]: ok, so CPAN and stuff :)
09:25:43 [sschaffert]: so we just wait for INFRA
09:25:44 [Wikier]: I'm trying to use maven as build system, in order to allow 
automatically publication of javadoc, rest api and so on
09:25:51 [Wikier]: yes
09:26:13 [Wikier]: I'd try to talk with them at #asfinfra
09:26:26 [sschaffert]: ok, once we have the mvn site, what change is needed 
from INFRA? we won't be using CMS then, true?
09:26:28 [Wikier]: hopefully to get something online early next week
09:27:04 [sschaffert]: this would be good, yes - at least so that people can 
see something, and we can maybe for the moment refer them to the old LMF page
09:27:12 [Wikier]: no, the cms supports different build systems
09:27:29 [Wikier]: maven is one
09:27:42 [sschaffert]: ah ok, so I suggest we use maven then instead of Perl :)
09:27:58 [Wikier]: so now the site is a maven artifact, where we could try to 
refer to the actual artifacts and publish technical doc there
09:28:41 [sschaffert]: so it would be possible to setup a Maven project for the 
website that generates the site out of the artifacts in the repository?
09:28:42 [Wikier]: what it should generate is more or less this: 
http://s.apache.org/marmotta
09:28:57 [Wikier]: that's my idea, yes
09:29:05 [Wikier]: let's see if we can do it
09:29:10 [jfrank]: so the workflow will be: have the source-files in git -> 
publish to svn -> run "mvn site" on to publish?
09:29:25 [Wikier]: not really
09:29:40 [sschaffert]: more like a separate project in SVN
09:29:48 [sschaffert]: that refers to the artifacts in the Maven repo
09:29:55 [Wikier]: I'd prefer being able to refer from the site directly the 
artifact, so no manual dump to svn would be needed
09:30:10 [Wikier]: let's see...
09:30:25 [sschaffert]: if this is possible, great - but not sure if mvn site 
supports such a thing
09:30:34 [jfrank]: so we need "-site" artifacts for the marmotta modules
09:30:48 [nandana]: take a look at how Apache maven project does this. I think 
they follow the same approach.
09:30:57 [sschaffert]: good suggestion
09:31:26 [sschaffert]: #action look how Apache Maven generates its site from 
artifacts
09:31:41 [Wikier]: exactly
09:31:50 [Wikier]: I'm trying to follow such approach
09:32:03 [sschaffert]: +1
09:32:18 [sschaffert]: so what is (in summary) the timeline?
09:32:26 [Wikier]: draft online early next week
09:32:50 [Wikier]: proper web sitr ecosystem during February
09:32:59 [sschaffert]: ok
09:33:03 [Wikier]: s/sitr/site
09:33:26 [sschaffert]: anything more regarding the website?
09:33:36 [Wikier]: no
09:33:41 [sschaffert]: so next topic


# 4. logo #
09:33:58 [sschaffert]: ok
09:34:11 [Wikier]: early this morning I send the proposal to [email protected]
09:34:12 [sschaffert]: latest version of logo is attached to the issue tracker
09:34:18 [Wikier]: awaiting for the feedback
09:34:29 [sschaffert]: #link https://issues.apache.org/jira/browse/MARMOTTA-5
09:34:33 [tkurz]: which one is the latest?
09:34:41 [sschaffert]: version 7
09:34:42 [Wikier]: the tarball
09:34:58 [tkurz]: okay. so lets see what happens
09:34:59 [sschaffert]: at the issue there is a ZIP file with pngs and a PDF for 
vector graphics
09:35:05 [sschaffert]: as soon as we get confirmation
09:35:11 [sschaffert]: we need to tell our designer
09:35:20 [tkurz]: is there a formal way to donate the logo do ASF?
09:35:20 [sschaffert]: because she will make improved versions for different 
sizes
09:35:56 [nandana]: I think when you attach it to the JIRA and say that you are 
willing to license it under AL2
09:35:57 [sschaffert]: this needs to be figured out, but maybe we can 
contribute it as part of the sourc code?
09:36:03 [Wikier]: plus the "Powered By..." logos
09:36:11 [sschaffert]: ok
09:36:14 [sschaffert]: we need to decide on the following things:
09:36:18 [Wikier]: images are not code
09:36:28 [Wikier]: ASF would need too register as trademark
09:36:34 [sschaffert]: but maybe the SVG files can be seen as code :)
09:36:41 [Wikier]: but not sure if during incubation of when graduating
09:36:41 [nandana]: the donation should be ok, I think (just my opinion, I 
think someone will confirm in general)
09:36:56 [nandana]: but you need to talk to trademarks@
09:37:03 [sschaffert]: trademarking is yet another issue
09:37:03 [nandana]: to make sure it is suitable
09:37:12 [nandana]: the use of the logo and stuff
09:37:34 [Wikier]: exactly
09:37:42 [nandana]: "Projects may may choose to use the Apache feather in their 
logo if they wish, but if they do so they should contact trademarks@ to 
determine the best format of the feather to use."
09:37:57 [nandana]: http://www.apache.org/foundation/marks/pmcs.html#graphics
09:37:57 [jfrank]: so we should do this before deciding on the logo
09:38:13 [sschaffert]: yes, we read the page
09:38:43 [nandana]: what I am not sure is weather all the trademark guidelines 
for TLPs also apply for podlings
09:38:57 [nandana]: like the use of TM etc.
09:38:59 [sschaffert]: well, in any case it does not hurt
09:39:12 [sschaffert]: to apply as much of it as possible
09:39:34 [sschaffert]: so things to do:
09:39:36 [nandana]: yeah, and I think the best one to advice on this is 
trademarks@
09:39:42 [jfrank]: well, we are definitely planning to be a TLP some day... ;-)
09:39:44 [sschaffert]: - contact trademarks@ refarding the use
09:40:12 [sschaffert]: - decide on the different sizes and formats we need (so 
our designer can create different versions)
09:40:57 [sschaffert]: - decide on any other uses of the brand we want to have 
(e.g. the marmot alone, or different background colours)
09:41:57 [sschaffert]: sergio, can you contact trademarks@ or should I do it?
09:41:57 [Wikier]: I'd suggest to await the feedback from IPMC
09:42:04 [sschaffert]: ok sure
09:42:05 [tkurz]: okay. and then we donate the whole set to ASF
09:42:14 [Wikier]: but yes, meanwhile I can contact to trademarks@
09:42:42 [Wikier]: #action contact to trademarks@ regarding the logo proposal
09:42:57 [sschaffert]: #action decision on different sizes and formats for the 
logo
09:43:27 [sschaffert]: #action decision on other uses of the brand in the 
software and on the website
09:43:42 [Wikier]: good!
09:43:44 [Wikier]: next topic?


# 5. administration interface #
09:44:06 [sschaffert]: Thomas, short report? :)
09:44:27 [tkurz]: yes. I started to simplify the layout
09:44:52 [tkurz]: it is easy to use for both marmotta and lmf
09:45:21 [tkurz]: furthermore i started to restructure the module properties 
files
09:45:51 [tkurz]: so the new admin interface should be finished today
09:46:06 [sschaffert]: side remark: when switching to OSGi, the 
module.properties will anyways probably be solved differently
09:46:19 [sschaffert]: but for now this is good
09:46:42 [tkurz]: okay. it is not a big issue (won't replace it by XML or 
something)
09:46:57 [tkurz]: just small changes
09:47:19 [Wikier]: this is a mid-term task, indeed
09:47:21 [sschaffert]: any place to see the new layout?
09:47:49 [sschaffert]: (maybe attachment to JIRA issue?)
09:48:05 [tkurz]: not yet. I'll attach a screenshot
09:48:19 [sschaffert]: +1
09:48:42 [Wikier]: +1
09:48:44 [sschaffert]: Christian also volunteered to continue working on the 
interface once he finished moving to Karlsruhe
09:48:57 [tkurz]: #action attach screenshot of new admin interface the JIRA 
issue
09:48:57 [sschaffert]: so maybe we should keep him in the loop
09:49:42 [sschaffert]: ok, any more issues regarding this topic?
09:49:57 [westei]: If he plans to contribute he should become a comitter
09:50:13 [sschaffert]: @westei: yes, but this still needs to be seen
09:50:50 [sschaffert]: next topic?
09:51:05 [Wikier]: anything else?
09:51:12 [Wikier]: umm, yes


# 6. w3c #
09:51:42 [Wikier]: there are some tasks regarding this
09:51:57 [Wikier]: the first, I plan to start to join the telco next monday
09:52:04 [sschaffert]: +1
09:52:05 [Wikier]: and start to involucrate myself a bit more
09:52:12 [tkurz]: +1
09:52:29 [Wikier]: in addtion, tkurz and me we'll start to explore the 
implementation of LDP-C
09:52:42 [sschaffert]: good
09:52:42 [Wikier]: I have several questions regarding this
09:52:44 [sschaffert]: my idea here is
09:52:51 [sschaffert]: once we switch to OSGi
09:53:04 [sschaffert]: to allow configuring different backend repositories for 
different containers
09:53:05 [Wikier]: imho there are like small graphs with some meta-properties 
and restrictions
09:53:19 [Wikier]: but in the mailing list there are some open discussions
09:53:49 [Wikier]: aggregation vs.
09:53:57 [Wikier]: and so on
09:54:05 [sschaffert]: there are many issues still unsolved
09:54:21 [sschaffert]: the basic question here is: what are containers meant for
09:54:21 [Wikier]: yes
09:54:28 [Wikier]: maybe nandana knows a bit more about it
09:54:45 [Wikier]: but I'd like to open such discussion in the following days
09:54:52 [Wikier]: they have some use cases
09:54:58 [sschaffert]: otherwise, try finding out in the telco
09:55:15 [Wikier]: #link 
https://dvcs.w3.org/hg/ldpwg/raw-file/default/ldp-ucr.html
09:55:49 [Wikier]: but we'll do deeper on it in the following days/weeks
09:55:52 [sschaffert]: so what is it you need from us until monday?
09:55:57 [tkurz]: okay. but you are not much wiser after reading this ;) too 
many open issues
09:56:04 [Wikier]: nothing
09:56:27 [sschaffert]: the use cases are very high-level
09:56:27 [Wikier]: just reading a bit more the background
09:56:34 [sschaffert]: don't tell much about concrete technical requirements
09:56:42 [tkurz]: I have some questions regarding the draft. maybe you can ask 
it during the telco?
09:56:44 [sschaffert]: or if containers are actually needed :)
09:57:19 [Wikier]: tkurz: good point: how to proceed with the communication and 
the wg
09:57:42 [Wikier]: #action create a jira to track the communication with the 
ldp wg
09:57:59 [sschaffert]: I have some issues too, but they are maybe more for 
future discussions
09:58:12 [Wikier]: ok
09:58:14 [sschaffert]: can I briefly summarize?
09:58:28 [Wikier]: just to point that we need to invest effort on such
09:58:29 [tkurz]: maybe we should start a discussion on the mailing list, so 
you can get the questions for the wg from there
09:58:43 [Wikier]: +1
09:58:50 [nandana]: +1
09:58:58 [sschaffert]: (1) update protocol - what to send in a POST/PUT/PATCH 
request
09:58:59 [Wikier]: #action start a discussion at the mailing list about ldp, 
specially ldp-c
09:59:14 [sschaffert]: (2) changeset representation format (e.g. the one of 
Talis)
09:59:22 [nandana]: the WG is also still agreeing on the stuff and open for 
proposals
09:59:37 [sschaffert]: (3) support for Memento protocol for versioning/snapshots
10:00:05 [sschaffert]: yes, and this is very good, so we can suggest things 
from technical experience
10:00:05 [jfrank]: is versioning/snapshots part of ldp?
10:00:14 [Wikier]: nandana: exactly, but I'd like to have a better overview of 
the current agreements before doing more proposal
10:00:15 [sschaffert]: no, but it could be a contribution/extension
10:00:43 [Wikier]: exactly
10:00:43 [sschaffert]: once you have updates, you will inevitably face the 
problem of logging, versioning, snapshots
10:00:44 [tkurz]: @jakob: it must be part of the http part
10:00:52 [tkurz]: or an extension
10:01:13 [sschaffert]: a concrete suggestion could be to add the Memento 
protocol as an extension
10:01:36 [sschaffert]: but what I would also like to have is "Permalinks" to 
"certified" snapshots
10:01:38 [jfrank]: so it's not yet in the ldp, but we'll try to bring it in.
10:01:58 [tkurz]: yes
10:02:13 [Wikier]: in addition, yes
10:02:28 [sschaffert]: but maybe now is not the time to discuss this all, maybe 
we should first get an overview what they are currently up to
10:02:28 [Wikier]: but I'd prefer to focus now of the main goals of ldp
10:02:44 [Wikier]: and provide feedback from a real implementation point of view
10:02:50 [sschaffert]: yes
10:03:06 [jfrank]: for that we should start with a "real implementation" (in 
marmotta)
10:03:13 [sschaffert]: so this leaves mainly the topic of the patches/updates 
and the topic of the changeset representation
10:03:13 [Wikier]: so we'll further discuss this in the upcoming weeks
10:03:22 [tkurz]: okay.any other issue for today?


# 7. aob #
10:03:43 [sschaffert]: #action continue discussion on LDP issues and 
contributions from Marmotta
10:03:50 [Wikier]: anything else?
10:04:00 [sschaffert]: not for today, coding is waiting :)
10:04:05 [Wikier]: 1 hour....
10:04:13 [Wikier]: coding is waiting, right!
10:04:14 [sschaffert]: needs to be more efficient in the future
10:04:20 [sschaffert]: 30 minutes max
10:04:29 [Wikier]: maybe providing an agenda in advance or so
10:04:38 [Wikier]: I'd say 15-20
10:04:43 [Wikier]: xD
10:04:50 [sschaffert]: yes :)
10:04:58 [Wikier]: well, thanks all for your time
10:05:05 [sschaffert]: thanks also
10:05:13 [sschaffert]: you send the minutes to dev@...?
10:05:28 [sschaffert]: uh oh
10:06:05 [sschaffert]: this was the wrong summary
10:06:20 [tkurz]: the one from yesterday
10:06:35 [sschaffert]: this meeting apparently had an error
10:06:35 [Wikier]: yes
10:06:36 [sschaffert]: -> infra?
10:06:43 [Wikier]: I'll report it to infra
10:06:50 [westei]: 
http://wilderness.apache.org/archives/apachemarmotta-18_01_2013-7848.html works 
now
10:06:58 [sschaffert]: I leave the window open just to have the log :)
10:07:13 [jfrank]: only the agenda
10:07:15 [westei]: you are right
10:07:15 [jfrank]: and not complete
10:07:28 [Wikier]: I guess we did something wrong...
10:07:52 [Wikier]: sorry
10:07:58 [Wikier]: I'll talk with infra
10:08:13 [tkurz]: we are still learning ;)
10:08:45 [Wikier]: yes :-)
10:09:05 [tkurz]: what about next chat?
10:09:07 [tkurz]: monday?
10:09:08 [Wikier]: nandana, fchrist: probably you would need to accept the 
email from ASFBot
10:09:13 [sschaffert]: monday 10:00 CET
10:09:20 [nandana]: yeah
10:09:28 [sschaffert]: maybe reject it!
10:09:29 [Wikier]: xD
10:09:35 [Wikier]: no
10:09:35 [nandana]: but at the moment it contains the same content as 
http://wilderness.apache.org/archives/apachemarmotta-18_01_2013-8856.txt
10:09:35 [sschaffert]: it is the old meeting summary, after all!
10:09:52 [nandana]: so does not make sense to send it to the dev list, right ?
10:10:05 [Wikier]: maybe not, right
10:10:05 [sschaffert]: no
10:10:14 [nandana]: but we can approve the address for later on
10:10:20 [sschaffert]: maybe sergio can try again
10:10:29 [sschaffert]: maybe now there is the updated version
10:11:19 [sschaffert]: ok, so happy coding :)
10:11:27 [Wikier]: I'll discuss it with infra, and send something later
10:12:07 [westei]: Wikier: I saved the text in a text file
10:12:18 [westei]: So if needed I can provide it later on
10:13:20 [Wikier]: I also have it in my client log, np, thx
10:13:54 [sschaffert]: if there are further issues, I'll hang around here
10:20:23 [Humbedooh]: terribly sorry for the inconvenience - it seems you had 
exhausted the bot's number of bullet points in your meeting, something I did 
not foresee
10:20:40 [Humbedooh]: do you still have a log of what went on? if so, I can 
recreate the summary for you
10:24:52 [Humbedooh]: and I can also see that you sent out my bla  bla bla 
summary to dev@ by mistake - oh dear
10:27:44 [nandana]: Humbedooh: don't worry about that, it got stuck in the 
moderation and we rejected it :)
10:27:58 [Humbedooh]: cool
10:28:00 [Humbedooh]: okay, let me just fix the bot
10:29:27 [nandana]: Humbedooh: thanks !

Reply via email to