5 ms·
The design went from looking like 1997 to looking like 2007. You can still see the old design in the docs: https://www.postgresql.org/docs/10/static/index.html
by everdev 8y ago
The design went from looking like 1997 to looking like 2007.
You can still see the old design in the docs: https://www.postgresql.org/docs/10/static/index.html https://www.postgresql.org/docs/10/static/index.html
I appreciate the update, but it's interesting from a design perspective how many technically amazing products have dated website design. I do appreciate not trying to force a fashionable design template onto every site, but there are some really slick minimal ones that can be used that will bring the design up to the current decade at least.
- jrobn 8y agoThe documentation pages are all that matter, and quite frankly, they are great.
- ximeng 8y agoExcept that every Google search returns a result from a different random version number.
- dspillett 8y agoThey can't really control that as a number of external factors part into Google's rankings (i.e. links from elsewhere), other than to make navigation to the right version easy which they seem to have managed (in, my, admittedly limited, experience there is always a version selection in the page head and the links go where you'd expect).
- ximeng 8y agoThe above comment is one of my most upvoted ever. It's annoying to have to click twice to get to the latest version. Maybe they make it too easy to get to old versions - I don't think I've ever seen a link to Wikipedia that points to an old version of a page. Not saying this is an easy problem to solve, and the documentation is mostly pretty great, but this definitely bugs me and apparently a bunch of other people too.
- marcosdumay 8y agoOld Wikipedia pages are just dated and not useful as a starting point. Documentation for old versions of Postgres are the most up to date page for anybody using that version. I am one of the people that gave you an upvote, but this is not a fair comparison at all.
- ximeng 8y agoNot aiming to be unfair to postgres - it's not an easy problem and the documentation is great, and it is different to Wikipedia where old content is generally objectively worse. I guess for postgres 80% of people would be happy with always landing on the current version, and making an extra click or adding a version to their search if they really need documentation for older versions. Maybe adding a brief version history at the bottom of documentation pages would reduce the need for old versions even further.
- kingbirdy 8y agoI think a solution might be to make links to the latest version redirect to a v/latest/ which redirects to the new latest version across upgrades. For each url v/1.2/docs if 1.2 is the new version, then it redirects to v/latest/docs and when copy-pasted to e.g. stack overflow, when someone comes along later when it's in version 1.5, it serves the v1.5 docs. So the latest Google crawl will see mostly v/latest/docs links, it will always serve those at the top of search results for docs, and then no matter what those search results at the latest version of docs.
- ximeng 8y agoThis would help! There are no direct links between the urls below which both have the same content. https://www.postgresql.org/docs/current/static/index.html https://www.postgresql.org/docs/current/static/index.html https://www.postgresql.org/docs/10/static/index.html https://www.postgresql.org/docs/10/static/index.html You have to click to another page and back again to change to the "current" link if you're on the "10" link.
- codetrotter 8y agoThey could straight up disallow the versioned pages in their robots.txt and for the current version they could have a full copy of the docs without version number in the URL that instead has .../current/... and those pages would be allowed for crawlers to index. Then people could jump to older versions manually from there if they needed to instead of everyone having to endure the land-on-an-arbitrary-version-number-page annoyance. That really is my only problem with the postgresql docs though and that’s saying something. The content itself of the postgresql docs is of stellar quality.
- h1d 8y agoNot sure how exactly Google works but if you just block versioned pages out, won't they just not show up instead of somehow pulling the latest static URL in the same ranked position? Maybe better to specify canonical URL as the latest one.
- slavik81 8y agoBlocking the pages in robots.txt will remove them from the Wayback Machine. That sucks. Qt did that for their documentation when they changed ownership and moved their website. They didn't migrate the old documentation, either. This made updating broken links much more difficult because I couldn't easily read the pages they originally pointed to.
- xemoka 8y agoWould defining a canonical page, `<link rel="canonical" href="...">`, be the solution to this?
- jimktrains2 8y agoIt's not canonical though, it's for different versions, they're not the same page.
- majewsky 8y agoMaybe we need <link rel="preferred-search-result">
- jimktrains2 8y agoI wonder if this could be "done" or managed in some way using a sitemap or via google's site search admin (I forget its exact name). Using robots to disallow old versions seems like an awfully big hammer.
- gsich 8y agoThere is a version link at the top.
- JoshMnem 8y agoYou can add a version number to the query to get a specific version: https://www.google.com/search?q=postgres+functions+9.1 https://www.google.com/search?q=postgres+functions+9.1
- karmakaze 8y agoI just follow a near-version link and edit the version number in the URL, then presto.
- gsich 8y agoThis was the first thing I checked, thankfully they stayed the same.
- siruncledrew 8y agoIdk, the examples are really the best part of the documentation to me personally, but they are all the way at the bottom of the page. The description and options sections just remind me of man pages. I would prefer more of a readthedocs style.
- beefhash 8y agoJust hit the "End" key to scroll to the bottom? Incidentally, UNIX/Linux man pages also put the examples section near the end.
- bachmeier 8y ago> it's interesting from a design perspective how many technically amazing products have dated website design It's interesting from a user perspective how many nice websites have been ruined in an attempt to not look dated.
- twodave 8y agoThis, in my opinion, is a little harsh.
- ArlenBales 8y agoIt's also misleading because responsive design was nonexistent in 2007, and this site looks quite good in mobile.
- tabtab 8y agoRe: You can still see the old design in... So we can see the PregreSQL and the PostgreSQL ;-)
- ken 8y agoI find it quite convenient that all of the best database systems I've used (Postgres, SQLite, Redis) also have the most boring webpages, and the worst databases have the most trendy webpages. I could have saved a lot of time, over the past 20 years, if I'd simply followed the rule: Is this the most boring webpage I've seen all week? Yes => it's going to be a good database.
- btown 8y agoOne important characteristic of a database, though, is that it remains commercially supported and has developer continuity. To that end, competence in modern enterprise sales can be correlated with modern web design, no? CockroachDB, for instance, is able to simultaneously focus on Jepsen-style distributed systems verification and agile administrative interfaces, whose design aesthetic is reflected in the website. And there are plenty of unsupported academia-made databases that are essentially abandonware. I would not advocate for overreliance on this heuristic.
- mistrial9 8y ago> unsupported academia-made databases that are essentially abandonware whoa - you are setting yourself up, if that is any reference to Postgres..
- btown 8y agoWasn't referring to Postgres at all (it's the most solid RDBMS I've worked with). And I may have been unfair to other databases - was just saying that like most tech products, there are varying degrees of business-readiness in the database space.
- hitekker 8y agoIn my view, you weren't referring to Postgres. Rather, the cognitive dissonance brought on by the shortcomings of the redesign are making some supporters extra-sensitive to criticism.
- 8y ago
- tim333 8y agoMaybe they could have all the elements load slowly with Ajax for that modern feel.
- TeMPOraL 8y ago> The design went from looking like 1997 to looking like 2007. Good. I hope someone in the team was saying, "The line must be drawn here! This far, no further!" - because it's some time after 2007 when the web went entirely to shit.
- cup-of-tea 8y agoAnyone in a position to care about postgresql doesn't care what year they think the website looks like.
- Someone 8y agoSites aimed at the enterprise message “trustworthy” by looking distinguished and a bit grey along the temples. Sites for databases and file systems doubly so. You don’t want to trust your data to a kid, no matter how smart. Similarly, a site for a cheap grocery store must look cheap. Customers shouldn’t think millions were spent on it. Examples: https://aldi.com https://aldi.com, https://www.lidl.com https://www.lidl.com