perf.haskell.org’s list of branches got more useful
Hi, the list of branches on https://perf.haskell.org/ghc got more useful. It now shows the difference from the branch head to the merge base¹, which should roughly exhibit the performance change that would be caused by merging the branch into master. If the branch head or the merge base have not been measured (e.g. because the build failed), the display might be a bit confusing. Same with branches that often merge or cherry-pick from master. But in the common case of a linear, possibly rebased, feature branch, it should work fine. Greetings, Joachim ¹ Which, I think, is the youngest commit also reachable from master. Or something like that. Consult "man git merge-base" if you want a little mind teaser. -- Joachim “nomeata” Breitner mail@joachim-breitner.de • https://www.joachim-breitner.de/ XMPP: nomeata@joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F Debian Developer: nomeata@debian.org
Joachim Fabulous. But your email will soon disappear into the mists of time, and the page itself does not convey the same information. Is there a wiki page, kept up to date, that explains exactly how to read the performance results? Simon | -----Original Message----- | From: ghc-devs [mailto:ghc-devs-bounces@haskell.org] On Behalf Of | Joachim Breitner | Sent: 23 April 2016 20:41 | To: ghc-devs@haskell.org | Subject: perf.haskell.org’s list of branches got more useful | | Hi, | | the list of branches | on https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fper | f.haskell.org%2fghc&data=01%7c01%7csimonpj%40064d.mgd.microsoft.com%7ce | c0ca8588956494efa3308d36baf36f5%7c72f988bf86f141af91ab2d7cd011db47%7c1& | sdata=Rk2Y38BjXeE9Et8yxiuOdpiJ3PSMn9RdIxonRksxTc4%3d got more useful. | It now shows the difference from the branch head to the merge base¹, | which should roughly exhibit the performance change that would be | caused by merging the branch into master. | | If the branch head or the merge base have not been measured (e.g. | because the build failed), the display might be a bit confusing. Same | with branches that often merge or cherry-pick from master. But in the | common case of a linear, possibly rebased, feature branch, it should | work fine. | | Greetings, | Joachim | | ¹ Which, I think, is the youngest commit also reachable from master. Or | something like that. Consult "man git merge-base" if you want a little | mind teaser. | | -- | Joachim “nomeata” Breitner | mail@joachim-breitner.de • | https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fwww.jo | achim- | breitner.de%2f&data=01%7c01%7csimonpj%40064d.mgd.microsoft.com%7cec0ca8 | 588956494efa3308d36baf36f5%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata | =ya%2bnaZWlbasw8BSIICImvWjh1x%2fb6%2bTsHJuKRGzu2Rg%3d | XMPP: nomeata@joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F | Debian Developer: nomeata@debian.org
Dear Simon, Am Montag, den 25.04.2016, 09:32 +0000 schrieb Simon Peyton Jones:
But your email will soon disappear into the mists of time, and the page itself does not convey the same information.
Is there a wiki page, kept up to date, that explains exactly how to read the performance results?
no, there is not, but I have doubts it will stay up-to-date. Also, the performance dashboard is not GHC specific, so the GHC trac is an odd place for documentation. Generally, UI that needs extensive documentation is bad UI. So if the interface is confusing, gipeda needs to be improved (and may it be with “title” tags that explain the fields). If there are concrete issues with the page, let me know, and I’ll try to improve the interface. Greetings, Joachim -- Joachim “nomeata” Breitner mail@joachim-breitner.de • https://www.joachim-breitner.de/ XMPP: nomeata@joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F Debian Developer: nomeata@debian.org
| Generally, UI that needs extensive documentation is bad UI. | So if the interface is confusing, gipeda needs to be improved (and may | it be with “title” tags that explain the fields). If there are concrete | issues with the page, let me know, and I’ll try to improve the | interface. Well that's a great aspiration, but it is a hard one to deliver on. I am completely flummoxed by almost every element of the current page. (And not just this UI, but many others.) * Nothing says what the baseline is; your email specifically says that you have changed the baseline To take the first line of the "Branches" section * What does "2 days ago" mean? * What does "3 commit to cb300e" mean? * What does 276 mean * What does the little histogram mean? * What does 0+0- mean? I'm short, I'm lost. I need a little page with an example and explanations of each element. Simon | -----Original Message----- | From: ghc-devs [mailto:ghc-devs-bounces@haskell.org] On Behalf Of | Joachim Breitner | Sent: 25 April 2016 11:34 | To: ghc-devs@haskell.org | Subject: Re: perf.haskell.org’s list of branches got more useful | | Dear Simon, | | | Am Montag, den 25.04.2016, 09:32 +0000 schrieb Simon Peyton Jones: | > But your email will soon disappear into the mists of time, and the | > page itself does not convey the same information. | > | > Is there a wiki page, kept up to date, that explains exactly how to | > read the performance results? | | no, there is not, but I have doubts it will stay up-to-date. Also, the | performance dashboard is not GHC specific, so the GHC trac is an odd | place for documentation. | | Generally, UI that needs extensive documentation is bad UI. | So if the interface is confusing, gipeda needs to be improved (and may | it be with “title” tags that explain the fields). If there are concrete | issues with the page, let me know, and I’ll try to improve the | interface. | | Greetings, | Joachim | | | -- | Joachim “nomeata” Breitner | mail@joachim-breitner.de • | https://na01.safelinks.protection.outlook.com/?url=https%3a%2f%2fwww.jo | achim- | breitner.de%2f&data=01%7c01%7csimonpj%40064d.mgd.microsoft.com%7c84e8b9 | fcdb4b4c19445808d36cf528d9%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata | =bDjE6VphGJnU1mqFdjWKkU2dH4o5554xtdx%2fl%2fLreoM%3d | XMPP: nomeata@joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F | Debian Developer: nomeata@debian.org
Hi, Am Montag, den 25.04.2016, 13:54 +0000 schrieb Simon Peyton Jones:
Generally, UI that needs extensive documentation is bad UI. So if the interface is confusing, gipeda needs to be improved (and may it be with “title” tags that explain the fields). If there are concrete issues with the page, let me know, and I’ll try to improve the interface.
Well that's a great aspiration, but it is a hard one to deliver on. I am completely flummoxed by almost every element of the current page. (And not just this UI, but many others.)
And I am repeatedly flummoxed by your vocabularly; always a joy :-)
* Nothing says what the baseline is; your email specifically says that you have changed the baseline
To take the first line of the "Branches" section
[..]
I'm short, I'm lost. I need a little page with an example and explanations of each element.
Hmm, I see your point. But since I did not see it originally, maybe I’m not the right person to describe the UI – after all, I’m heavily biased by having it created in the first place. Are there volunteers that that would enjoy creating a user-oriented documentation for gipeda? (Probably not in the form of a detached wiki page, but a help page linked from https://perf.haskell.org/ghc directly.) Until then, allow me to briefly answer your questions:
* What does "2 days ago" mean?
The age of the latest commit to the branch.
* What does "3 commit to cb300e" mean?
That this branch contains 3 commits relative to master, and that the final commit on the branch is cb300e. These are two links with tooltips: The first one says „Compare against merge base“, which will take you to a comparison page between the base commit and the latest commit on the branch. The second one says „Show branch tip“, will will take you to that revisions’ usual results view, which is relative to its parent.
* What does 276 mean
Number of benchmarks taken. I’m considering to remove this bit of information here, it is not really interesting.
* What does the little histogram mean * What does 0+0- mean?
Number of improvements (+) and regressions (-). In the per-branch-list is against the merge base, in the commit list it is against the commit’s parent. If you hover, you’ll see a list of changes (if there are any). Greetings, Joachim -- Joachim “nomeata” Breitner mail@joachim-breitner.de • https://www.joachim-breitner.de/ XMPP: nomeata@joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F Debian Developer: nomeata@debian.org
On 04/26/2016 01:09 PM, Joachim Breitner wrote:
Hi,
* What does "2 days ago" mean?
The age of the latest commit to the branch.
FWIW, I think adding a few simple table headings might help enormously. "Last commit", "?", "Branch", "Last commit message", "Diffstat". (or something like that.)
Hi, Am Dienstag, den 26.04.2016, 16:57 +0200 schrieb Bardur Arantsson:
On 04/26/2016 01:09 PM, Joachim Breitner wrote:
* What does "2 days ago" mean? The age of the latest commit to the branch.
FWIW, I think adding a few simple table headings might help enormously.
"Last commit", "?", "Branch", "Last commit message", "Diffstat".
(or something like that.)
good idea; although often with tables space is thight, and narrow columns don’t allow for wide enough headers. Greetings, Joachim -- Joachim “nomeata” Breitner mail@joachim-breitner.de • https://www.joachim-breitner.de/ XMPP: nomeata@joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F Debian Developer: nomeata@debian.org
Joachim Breitner
FWIW, I think adding a few simple table headings might help enormously.
"Last commit", "?", "Branch", "Last commit message", "Diffstat".
(or something like that.)
good idea; although often with tables space is thight, and narrow columns don’t allow for wide enough headers.
Just a random wild idea -- rotate text by approximately 90 degrees? -- с уважениeм / respectfully, Косырев Сергей
On 04/26/2016 11:11 PM, Joachim Breitner wrote:
Hi,
Am Dienstag, den 26.04.2016, 16:57 +0200 schrieb Bardur Arantsson:
On 04/26/2016 01:09 PM, Joachim Breitner wrote:
* What does "2 days ago" mean? The age of the latest commit to the branch.
FWIW, I think adding a few simple table headings might help enormously.
"Last commit", "?", "Branch", "Last commit message", "Diffstat".
(or something like that.)
good idea; although often with tables space is thight, and narrow columns don’t allow for wide enough headers.
You can always shorten the text or use abbreviations and then add hover text with the 'full name' of the column. Mouse hover text isn't *great* UX-wise, but probably better than nothing. It's generally a little less 'discoverable' even if people are generally used to the idea of hover tooltips. Additionnaly, I'm not sure how well this will work for blind people who use screen readers. Regards,
Hi, Am Mittwoch, den 27.04.2016, 05:42 +0200 schrieb Bardur Arantsson:
You can always shorten the text or use abbreviations and then add hover text with the 'full name' of the column. Mouse hover text isn't *great* UX-wise, but probably better than nothing. It's generally a little less 'discoverable' even if people are generally used to the idea of hover tooltips. Additionnaly, I'm not sure how well this will work for blind people who use screen readers.
I believe this is getting off topic on this list; I invite you to continue brainstorming on https://github.com/nomeata/gipeda/issues/37 Greetings, Joachim -- Joachim “nomeata” Breitner mail@joachim-breitner.de • https://www.joachim-breitner.de/ XMPP: nomeata@joachim-breitner.de • OpenPGP-Key: 0xF0FBF51F Debian Developer: nomeata@debian.org
participants (4)
-
Bardur Arantsson -
Joachim Breitner -
Kosyrev Serge -
Simon Peyton Jones