I think both tools have their merits and drawbacks
What I like about Jira:
- It has ample room and configuration for issue metadata and
customizable workflows and in general a deep feature set
- It has user roles, PMC members can see security issues that are hidden
from the world...
- I've used it for almost 20 years so It's familiar to me.
- It's hosted at the ASF by the ASF so nobody but the ASF can determine
access or hold it hostage (I think, correct me if I'm wrong and we're now
using atlassian cloud versions).
What Ii do not like about Jira
- They have had LONG standing issues with text and visual mode not
round-tripping (switching between them alters the text and often destroys
formatting) which is something even cheap blogging software usually gets
right.... this is EXTREMELY FRUSTRATING at times where the proper name of
something in code includes an underscore. Especially bad is the fact that
the do provide a way to escape things like underscores but the transition
between visual and text destroys that escaping, making it useless, and if
you carefully set up the escapes in text mode, one small edit by someone
else (perhaps fixing a typo) in visual mode destroys all your hard work!
GRRRR... And of course text mode is sometimes hard to predict so working in
text mode with any non-trivial formatting that you may wind up re-editing
several times which at apache sends multiple emails to the list...
usability nightmare.
- They switched to a default search result layout that wasn't a sortable
table/list. This irritates me because I never want to randomly fill 70% of
the screen with the top hit on a search and have almost no info about all
the other results. Typically I want to immediately sort by issue number to
find recent issues (or older issues depending). Even if text relevancy is
the important thing in my search, assuming the top hit is what I want is
poor.
- By default searches typed in the easily accessed search box run
against every project so then the very first thing I have to do is re-run
with a project filter. Maintaining a project context would be very helpful.
- UI can become cumbersome for filtering on issue fields with many
values (typeahead search presumes you know the name to start with).
- Sometimes slow.
What I like about Github
- It fixes the first and third issue I don't like about jira (edit round
trip and typically has a project context).
- UI updates without explicit refresh
- Generally nice look and feel
- Integration of github actions, pull requests, review, and code
repository is excellent.
- Closing issues via commit message is nice for small projects...
What I don't like about github.
- Very limited and no custom issue metadata
- arbitrary file attachments are not supported. Notably .patch is not
included in their list (GIF, JPEG, JPG, MOV, MP4, PNG, SVG, CSV, DOCX,
FODG, FODP, FODS, FODT, GZ, LOG, MD, ODF, ODG, ODP, ODS, ODT, PDF, PPTX,
TXT, XLS, XLSX or ZIP).
- Search interface for issues is the equivalent of the Jira JQL search
line, and requires learning their syntax for anything but the most basic,
and is:issue must be retained (and wastes space in the small text field) or
it suddenly starts finding non-issues items.
- No concept of workflow (without add/on or plugins).
- Closing issues via commit message is not good where you would want to
ensure review or have any sort of workflow
- It's owned by Microsoft, which while MUCH improved in recent years has
a horrible dark, evil past WRT open source and standards. Above allegations
regarding political banning of individuals is also very troubling and
unfortunately, increasingly relevant in the current global political
landscape.
From the lucene perspective I see some things we have now in Jira that I
don't see a way to maintain in github
- Security issues that are visible to PMC only (more common for solr,
but perhaps needed for lucene sometimes as well)
- Patch review based on an attached patch.
- Accepting patch files instead of pull requests...
- Contribution by folks who for some reason cannot use github (either
blocked by work or github politics or, unwilling to accept githubs
terms/privacy etc.)
- Document with each issue the Affected version and the fixed version.
- While one can create arbitrary labels, they are not segregated into
fields so we would have to put up with what is effectively a single field
for priority, component, and resolution
- No way to enforce that a resolution label is applied to the issue.
I feel that Github issues are simply lacking in depth and riding along on
the virtue of their integrations. I feel like their issue tracking
implementation is a lower priority sideline to their code repository (so
they can say they have it). On the flip side Jira has become hugely
entrenched in the industry and its profits are no-longer tied to innovation
or even usability... and it shows. So I am basically dissatisfied with both
(for most of the last 20 years I've been known to mutter about writing my
own issue tracker... I've not met one that didn't irritate me somehow,
though I haven't actually tried yet, because that's a LOT of work ;) ).
Given how many things I've listed, it's likely something's wrong or
misapprehended, so certainly speak up if I'm just unaware of something in
github.
In the end I don't feel like the benefits of github are worth giving up the
data quality and features that we do actually use in Jira, so I'm -0.9 on
moving to github.
-Gus
On Fri, May 6, 2022 at 11:11 AM Michael McCandless <
lucene@mikemccandless.com> wrote:
> On Thu, May 5, 2022 at 7:56 AM Robert Muir <rcmuir@gmail.com> wrote:
>
> As far as replies, in github I highlight the part of the thing i want
>> to reply to, and press 'r' key on my keyboard. it quotes it and
>> everything. Really convenient to me.
>>
>
> Whoa, thank you!! I had no idea GitHub has such extensive keyboard
> shortcuts (just type ? to see them all).
>
> Mike McCandless
>
> http://blog.mikemccandless.com
>
--
http://www.needhamsoftware.com (work)
http://www.the111shift.com (play)
What I like about Jira:
- It has ample room and configuration for issue metadata and
customizable workflows and in general a deep feature set
- It has user roles, PMC members can see security issues that are hidden
from the world...
- I've used it for almost 20 years so It's familiar to me.
- It's hosted at the ASF by the ASF so nobody but the ASF can determine
access or hold it hostage (I think, correct me if I'm wrong and we're now
using atlassian cloud versions).
What Ii do not like about Jira
- They have had LONG standing issues with text and visual mode not
round-tripping (switching between them alters the text and often destroys
formatting) which is something even cheap blogging software usually gets
right.... this is EXTREMELY FRUSTRATING at times where the proper name of
something in code includes an underscore. Especially bad is the fact that
the do provide a way to escape things like underscores but the transition
between visual and text destroys that escaping, making it useless, and if
you carefully set up the escapes in text mode, one small edit by someone
else (perhaps fixing a typo) in visual mode destroys all your hard work!
GRRRR... And of course text mode is sometimes hard to predict so working in
text mode with any non-trivial formatting that you may wind up re-editing
several times which at apache sends multiple emails to the list...
usability nightmare.
- They switched to a default search result layout that wasn't a sortable
table/list. This irritates me because I never want to randomly fill 70% of
the screen with the top hit on a search and have almost no info about all
the other results. Typically I want to immediately sort by issue number to
find recent issues (or older issues depending). Even if text relevancy is
the important thing in my search, assuming the top hit is what I want is
poor.
- By default searches typed in the easily accessed search box run
against every project so then the very first thing I have to do is re-run
with a project filter. Maintaining a project context would be very helpful.
- UI can become cumbersome for filtering on issue fields with many
values (typeahead search presumes you know the name to start with).
- Sometimes slow.
What I like about Github
- It fixes the first and third issue I don't like about jira (edit round
trip and typically has a project context).
- UI updates without explicit refresh
- Generally nice look and feel
- Integration of github actions, pull requests, review, and code
repository is excellent.
- Closing issues via commit message is nice for small projects...
What I don't like about github.
- Very limited and no custom issue metadata
- arbitrary file attachments are not supported. Notably .patch is not
included in their list (GIF, JPEG, JPG, MOV, MP4, PNG, SVG, CSV, DOCX,
FODG, FODP, FODS, FODT, GZ, LOG, MD, ODF, ODG, ODP, ODS, ODT, PDF, PPTX,
TXT, XLS, XLSX or ZIP).
- Search interface for issues is the equivalent of the Jira JQL search
line, and requires learning their syntax for anything but the most basic,
and is:issue must be retained (and wastes space in the small text field) or
it suddenly starts finding non-issues items.
- No concept of workflow (without add/on or plugins).
- Closing issues via commit message is not good where you would want to
ensure review or have any sort of workflow
- It's owned by Microsoft, which while MUCH improved in recent years has
a horrible dark, evil past WRT open source and standards. Above allegations
regarding political banning of individuals is also very troubling and
unfortunately, increasingly relevant in the current global political
landscape.
From the lucene perspective I see some things we have now in Jira that I
don't see a way to maintain in github
- Security issues that are visible to PMC only (more common for solr,
but perhaps needed for lucene sometimes as well)
- Patch review based on an attached patch.
- Accepting patch files instead of pull requests...
- Contribution by folks who for some reason cannot use github (either
blocked by work or github politics or, unwilling to accept githubs
terms/privacy etc.)
- Document with each issue the Affected version and the fixed version.
- While one can create arbitrary labels, they are not segregated into
fields so we would have to put up with what is effectively a single field
for priority, component, and resolution
- No way to enforce that a resolution label is applied to the issue.
I feel that Github issues are simply lacking in depth and riding along on
the virtue of their integrations. I feel like their issue tracking
implementation is a lower priority sideline to their code repository (so
they can say they have it). On the flip side Jira has become hugely
entrenched in the industry and its profits are no-longer tied to innovation
or even usability... and it shows. So I am basically dissatisfied with both
(for most of the last 20 years I've been known to mutter about writing my
own issue tracker... I've not met one that didn't irritate me somehow,
though I haven't actually tried yet, because that's a LOT of work ;) ).
Given how many things I've listed, it's likely something's wrong or
misapprehended, so certainly speak up if I'm just unaware of something in
github.
In the end I don't feel like the benefits of github are worth giving up the
data quality and features that we do actually use in Jira, so I'm -0.9 on
moving to github.
-Gus
On Fri, May 6, 2022 at 11:11 AM Michael McCandless <
lucene@mikemccandless.com> wrote:
> On Thu, May 5, 2022 at 7:56 AM Robert Muir <rcmuir@gmail.com> wrote:
>
> As far as replies, in github I highlight the part of the thing i want
>> to reply to, and press 'r' key on my keyboard. it quotes it and
>> everything. Really convenient to me.
>>
>
> Whoa, thank you!! I had no idea GitHub has such extensive keyboard
> shortcuts (just type ? to see them all).
>
> Mike McCandless
>
> http://blog.mikemccandless.com
>
--
http://www.needhamsoftware.com (work)
http://www.the111shift.com (play)