You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@iotdb.apache.org by Christofer Dutz <ch...@c-ware.de> on 2019/10/05 12:48:14 UTC
How about a summary of what you're all up to?
Hi all,
Again I noticed my Mail client showed me the usual picture … a 2 digit number of emails in the commit list, and 1 in the dev list.
Would be cool if you could sum up what the team is currently working on.
Are there any new names on the contributor side? You haven’t voted in anyone for quite some time
Chris
Re: How about a summary of what you're all up to?
Posted by Justin Mclean <ju...@classsoftware.com>.
Hi,
> overall my impression is that its getting better.
+1 to that and both what Julian and Chris have said. I think the next step is a little more transparency on how how people decided on what issues to work on.
Thanks,
Justin
Re: How about a summary of what you're all up to?
Posted by Julian Feinauer <j....@pragmaticminds.de>.
Hi,
overall my impression is that its getting better.
Initially it was way calmer here but now there are discussion threads or announcements about features from time to time.
So I totally agree with chris but I also think there is already a learning curve.
So keep it going guys : )
Julian
Am 06.10.19, 15:22 schrieb "Xiangdong Huang" <sa...@gmail.com>:
Hi,
+1 for more discussions, rather than just opening issues, working on them
and closing them.
Indeed, there are some discussions for some issues, but we can do more
about that.
If the issue is for a new feature, we can show the main idea about how to
implement it on the mailing list.
I summarize some new features/improvements we have solved or addressed in
September according to the mailing list, and github as following with 3
columns: Feature/imporvement description. status. Is there some discussions?
(The statistics may be not accurate. Some of the issues have been solved
but the status of the issues are not changed. Some discussions are on
github but I do not review the contents of related PRs.)
- Improvement for LRUCache. Not solved. no discussion.
- Reimplementation sync module. not solved. no discussion.
- improving spark-tsfile query for narrow table view. not solved. no
discussion.
- Spark-IoTDB connector. not solved. no discussion.
- maintain hadoop-tsfile module. not solved. no discussion (but the
contributor pointed out his main idea).
- load configuration file on-site. not solved. no discussion.
- connect IoTDB with python or C. not solved. no discussion but there are
some replies for that.
- use hive to query Tsfile data. solved. no discussion.
- switch Antlr3 to antlr4. Not solved. Some discussions (finally, the
performance using antlr4 is worse?).
- enrich session interfaces (using API rather than SQL to operate data) Not
solved. Some discussions.
- support saving data file into HDFS directly. Solved. Some discussions.
- make the result format of metadata query like "show timeseries"
consistent. solved. no discussion but the contributor keeps updating his
progress on the mailing list.
- Create metadata automatically in the insertion. not solved. Some
discussions.
- A simple tool to visualize logs. not solved. Some discussions.
- A new result set format. not solved. no discussion. Some discussions.
- Add bloom filters to TsFile. not solved. Some discussions.
- support retention policy (remove stale data automatically). not solved.
few discussion.
- Convert Path(String) to ID(Long). not solved. few discussions.
- support connection pool (e.g., Ali druid). not solved. few discussion.
- TsFile skeleton viewer tool. solved. few discussion.
- support External sort for non-sequential data. solved. some discussions
on github.
After reviewing these "discussions", I find that there are few mails for
discussing the idea of how to implement new features.. which I think is the
same important with why the feature is needed or how to define the feature.
In my personal opinion, in the next 3 month (and then 2019 is over...), I
think there are two important features, distributed version and relational
model view.
Best,
-----------------------------------
Xiangdong Huang
School of Software, Tsinghua University
黄向东
清华大学 软件学院
Christofer Dutz <ch...@c-ware.de> 于2019年10月5日周六 下午11:24写道：
> Hi Xiangdong,
>
> well a Maven release usually does exactly 2 commits.
> But it's something I noticed quite often. Two digit number of commits, 1-2
> emails on dev.
> So everyone is working hard on all sorts of stuff, but I see absolutely no
> coordination here.
> Is everything just people open Issues, and then people start working on
> these?
> Usually I would sort of expect some discussions about things.
>
> Chris
>
> Am 05.10.19, 17:12 schrieb "Xiangdong Huang" <sa...@gmail.com>:
>
> Hi,
>
> if the mails are sent today, I think most of them are because I am
> running
> ‘mvn release:prepare’ and ‘mvn release:rollback’, others are for fixing
> issues in RC1. All of them are on the rep/0.8 branch.
>
> Best,
>
> Christofer Dutz <ch...@c-ware.de>于2019年10月5日 周六下午8:48写道：
>
> > Hi all,
> >
> > Again I noticed my Mail client showed me the usual picture … a 2
> digit
> > number of emails in the commit list, and 1 in the dev list.
> > Would be cool if you could sum up what the team is currently working
> on.
> >
> > Are there any new names on the contributor side? You haven’t voted in
> > anyone for quite some time
> >
> > Chris
> >
> --
> -----------------------------------
> Xiangdong Huang
> School of Software, Tsinghua University
>
> 黄向东
> 清华大学 软件学院
>
>
>
Re: How about a summary of what you're all up to?
Posted by Xiangdong Huang <sa...@gmail.com>.
Hi,
+1 for more discussions, rather than just opening issues, working on them
and closing them.
Indeed, there are some discussions for some issues, but we can do more
about that.
If the issue is for a new feature, we can show the main idea about how to
implement it on the mailing list.
I summarize some new features/improvements we have solved or addressed in
September according to the mailing list, and github as following with 3
columns: Feature/imporvement description. status. Is there some discussions?
(The statistics may be not accurate. Some of the issues have been solved
but the status of the issues are not changed. Some discussions are on
github but I do not review the contents of related PRs.)
- Improvement for LRUCache. Not solved. no discussion.
- Reimplementation sync module. not solved. no discussion.
- improving spark-tsfile query for narrow table view. not solved. no
discussion.
- Spark-IoTDB connector. not solved. no discussion.
- maintain hadoop-tsfile module. not solved. no discussion (but the
contributor pointed out his main idea).
- load configuration file on-site. not solved. no discussion.
- connect IoTDB with python or C. not solved. no discussion but there are
some replies for that.
- use hive to query Tsfile data. solved. no discussion.
- switch Antlr3 to antlr4. Not solved. Some discussions (finally, the
performance using antlr4 is worse?).
- enrich session interfaces (using API rather than SQL to operate data) Not
solved. Some discussions.
- support saving data file into HDFS directly. Solved. Some discussions.
- make the result format of metadata query like "show timeseries"
consistent. solved. no discussion but the contributor keeps updating his
progress on the mailing list.
- Create metadata automatically in the insertion. not solved. Some
discussions.
- A simple tool to visualize logs. not solved. Some discussions.
- A new result set format. not solved. no discussion. Some discussions.
- Add bloom filters to TsFile. not solved. Some discussions.
- support retention policy (remove stale data automatically). not solved.
few discussion.
- Convert Path(String) to ID(Long). not solved. few discussions.
- support connection pool (e.g., Ali druid). not solved. few discussion.
- TsFile skeleton viewer tool. solved. few discussion.
- support External sort for non-sequential data. solved. some discussions
on github.
After reviewing these "discussions", I find that there are few mails for
discussing the idea of how to implement new features.. which I think is the
same important with why the feature is needed or how to define the feature.
In my personal opinion, in the next 3 month (and then 2019 is over...), I
think there are two important features, distributed version and relational
model view.
Best,
-----------------------------------
Xiangdong Huang
School of Software, Tsinghua University
黄向东
清华大学 软件学院
Christofer Dutz <ch...@c-ware.de> 于2019年10月5日周六 下午11:24写道：
> Hi Xiangdong,
>
> well a Maven release usually does exactly 2 commits.
> But it's something I noticed quite often. Two digit number of commits, 1-2
> emails on dev.
> So everyone is working hard on all sorts of stuff, but I see absolutely no
> coordination here.
> Is everything just people open Issues, and then people start working on
> these?
> Usually I would sort of expect some discussions about things.
>
> Chris
>
> Am 05.10.19, 17:12 schrieb "Xiangdong Huang" <sa...@gmail.com>:
>
> Hi,
>
> if the mails are sent today, I think most of them are because I am
> running
> ‘mvn release:prepare’ and ‘mvn release:rollback’, others are for fixing
> issues in RC1. All of them are on the rep/0.8 branch.
>
> Best,
>
> Christofer Dutz <ch...@c-ware.de>于2019年10月5日 周六下午8:48写道：
>
> > Hi all,
> >
> > Again I noticed my Mail client showed me the usual picture … a 2
> digit
> > number of emails in the commit list, and 1 in the dev list.
> > Would be cool if you could sum up what the team is currently working
> on.
> >
> > Are there any new names on the contributor side? You haven’t voted in
> > anyone for quite some time
> >
> > Chris
> >
> --
> -----------------------------------
> Xiangdong Huang
> School of Software, Tsinghua University
>
> 黄向东
> 清华大学 软件学院
>
>
>
Re: How about a summary of what you're all up to?
Posted by Christofer Dutz <ch...@c-ware.de>.
Hi Xiangdong,
well a Maven release usually does exactly 2 commits.
But it's something I noticed quite often. Two digit number of commits, 1-2 emails on dev.
So everyone is working hard on all sorts of stuff, but I see absolutely no coordination here.
Is everything just people open Issues, and then people start working on these?
Usually I would sort of expect some discussions about things.
Chris
Am 05.10.19, 17:12 schrieb "Xiangdong Huang" <sa...@gmail.com>:
Hi,
if the mails are sent today, I think most of them are because I am running
‘mvn release:prepare’ and ‘mvn release:rollback’, others are for fixing
issues in RC1. All of them are on the rep/0.8 branch.
Best,
Christofer Dutz <ch...@c-ware.de>于2019年10月5日 周六下午8:48写道：
> Hi all,
>
> Again I noticed my Mail client showed me the usual picture … a 2 digit
> number of emails in the commit list, and 1 in the dev list.
> Would be cool if you could sum up what the team is currently working on.
>
> Are there any new names on the contributor side? You haven’t voted in
> anyone for quite some time
>
> Chris
>
--
-----------------------------------
Xiangdong Huang
School of Software, Tsinghua University
黄向东
清华大学 软件学院
Re: How about a summary of what you're all up to?
Posted by Xiangdong Huang <sa...@gmail.com>.
Hi,
if the mails are sent today, I think most of them are because I am running
‘mvn release:prepare’ and ‘mvn release:rollback’, others are for fixing
issues in RC1. All of them are on the rep/0.8 branch.
Best,
Christofer Dutz <ch...@c-ware.de>于2019年10月5日 周六下午8:48写道：
> Hi all,
>
> Again I noticed my Mail client showed me the usual picture … a 2 digit
> number of emails in the commit list, and 1 in the dev list.
> Would be cool if you could sum up what the team is currently working on.
>
> Are there any new names on the contributor side? You haven’t voted in
> anyone for quite some time
>
> Chris
>
--
-----------------------------------
Xiangdong Huang
School of Software, Tsinghua University
黄向东
清华大学 软件学院