I've made some investigation on Trac tickets and realized that almost 23% of the tickets are duplicates. Is this type of tickets really a problem to people that have to analyze them (in Trac)? Some academic work have addressed this problem telling that it is very serious, thus i want to confirm with you too.
On Sun, Apr 6, 2008 at 11:13 PM, Yguaratã C. Cavalcanti
<yguar...@gmail.com> wrote: > I've made some investigation on Trac tickets and realized that almost 23% of > the tickets are duplicates. Is this type of tickets really a problem to > people that have to analyze them (in Trac)? Some academic work have > addressed this problem telling that it is very serious, thus i want to > confirm with you too.
Hi,
It's certainly very time-consuming dealing with duplicate tickets. If you provided a patch that implemented robust duplicate detection, I suspect you would make many new friends :)
Alec -- Evolution: Taking care of those too stupid to take care of themselves.
Although that won't help you detect duplicates when they're already entered in the system. However, the mechanism for searching for duplicates would probably be quite similar if you search for duplicates before or after entering them..
I encourage you to start working on such a feature =)
On Sun, Apr 6, 2008 at 11:26 PM, Alec Thomas <a...@swapoff.org> wrote:
> On Sun, Apr 6, 2008 at 11:13 PM, Yguaratã C. Cavalcanti > <yguar...@gmail.com> wrote: > > I've made some investigation on Trac tickets and realized that almost 23% of > > the tickets are duplicates. Is this type of tickets really a problem to > > people that have to analyze them (in Trac)? Some academic work have > > addressed this problem telling that it is very serious, thus i want to > > confirm with you too.
> Hi,
> It's certainly very time-consuming dealing with duplicate tickets. If > you provided a patch that implemented robust duplicate detection, I > suspect you would make many new friends :)
> Alec > -- > Evolution: Taking care of those too stupid to take care of themselves.
So do you think would be helpful some web-based tool to search in past bug reports, before entering them? If so, in this case we assume that people always search for past reports. :P
I'm discussing about it because we (RiSE - http://www.rise.com.br/research) are developing some kind of tool to search in past reports using techniques appropriate for this task. Moreover the idea is to enable project's community, like Trac, use the tool online.
Basically, the idea behind the tool is to have a interface where people could index the project's reports and search for them. Of course, robust search techniques, which are not provide currently by bug trackers, are being used to perform this task.
Will Trac reporters use it?
Cheers.
On Mon, Apr 7, 2008 at 9:57 AM, Erling Wegger Linde <erlin...@gmail.com> wrote:
> Although that won't help you detect duplicates when they're already > entered in the system. However, the mechanism for searching for > duplicates would probably be quite similar if you search for > duplicates before or after entering them..
> I encourage you to start working on such a feature =)
> On Sun, Apr 6, 2008 at 11:26 PM, Alec Thomas <a...@swapoff.org> wrote:
> > On Sun, Apr 6, 2008 at 11:13 PM, Yguaratã C. Cavalcanti > > <yguar...@gmail.com> wrote: > > > I've made some investigation on Trac tickets and realized that almost > 23% of > > > the tickets are duplicates. Is this type of tickets really a problem > to > > > people that have to analyze them (in Trac)? Some academic work have > > > addressed this problem telling that it is very serious, thus i want > to > > > confirm with you too.
> > Hi,
> > It's certainly very time-consuming dealing with duplicate tickets. If > > you provided a patch that implemented robust duplicate detection, I > > suspect you would make many new friends :)
> > Alec > > -- > > Evolution: Taking care of those too stupid to take care of themselves.
For it to be useful it would have to be integrated into the ticket reporting process. Having it "opt-in" would not be useful. Ideally as a Trac plugin, but a patch against Trac itself could be useful too.
I honestly don't think relying on an external web service is a good option. We had many problems with Akismet in the past, for example.
Others might be interested, but I wouldn't use it on Trac Hacks, sorry :\
On 4/10/08, Yguaratã C. Cavalcanti <yguar...@gmail.com> wrote:
> So do you think would be helpful some web-based tool to search in past bug > reports, before entering them? If so, in this case we assume that people > always search for past reports. :P
> I'm discussing about it because we (RiSE - http://www.rise.com.br/research) > are developing some kind of tool to search in past reports using techniques > appropriate for this task. Moreover the idea is to enable project's > community, like Trac, use the tool online.
> Basically, the idea behind the tool is to have a interface where people > could index the project's reports and search for them. Of course, robust > search techniques, which are not provide currently by bug trackers, are > being used to perform this task.
> Will Trac reporters use it?
> Cheers.
> On Mon, Apr 7, 2008 at 9:57 AM, Erling Wegger Linde <erlin...@gmail.com> > wrote:
> > Although that won't help you detect duplicates when they're already > > entered in the system. However, the mechanism for searching for > > duplicates would probably be quite similar if you search for > > duplicates before or after entering them..
> > I encourage you to start working on such a feature =)
> > On Sun, Apr 6, 2008 at 11:26 PM, Alec Thomas <a...@swapoff.org> wrote:
> > > On Sun, Apr 6, 2008 at 11:13 PM, Yguaratã C. Cavalcanti > > > <yguar...@gmail.com> wrote: > > > > I've made some investigation on Trac tickets and realized that almost > 23% of > > > > the tickets are duplicates. Is this type of tickets really a problem > to > > > > people that have to analyze them (in Trac)? Some academic work have > > > > addressed this problem telling that it is very serious, thus i want > to > > > > confirm with you too.
> > > Hi,
> > > It's certainly very time-consuming dealing with duplicate tickets. If > > > you provided a patch that implemented robust duplicate detection, I > > > suspect you would make many new friends :)
> > > Alec > > > -- > > > Evolution: Taking care of those too stupid to take care of themselves.
