98 comments

[ 3.1 ms ] story [ 151 ms ] thread
I was just wondering why a feature as simple as this one with lots of demands is still not implemented (or rejected). (6 years !!)
Pretty funny how an ambiguous date format ("1/5/10") led to two interpretations of how long ago this was (6 or 10 years ago).

To clarify: at the time of initially writing this comment, there were only two comments in this thread: the parent of this comment, which said it was 6 years ago; and another, which said it was 10 years ago. I don't know or care whether it was legitimate for either to come to the conclusion they did, I just thought it was funny, and that it was most likely due to the super-short date format inside the link. Can we move on please?

How can 1/5/10 be interpreted as 10 years ago? The only two interpretations I can think of are "January 5, 2010" and "May 1, 2010".
https://en.wikipedia.org/wiki/Date_format_by_country

Peer into the madness, and quake in fear.

Sorry, I understand that different locales use different orderings, and I scanned that chart, but I still don't see how "1/5/10" could have anything but the "10" refer to the year. Is there a country that uses one-digit years in their dates (unless referring to 1AD or 5AD)?
South Africa use the Year/Month/Day order, and if one assumes these strings were converted to integers and that the conversion implies any number of preceding zeros are ignored, then that could be 10th of May 2001
I use year/month/day order too (so does ISO) but we all pad out our years to 2 or 4 digits.

If the only way to interpret "1/5/10" as 2001 is to assume the formatting is hosed, then I guess one could also assume the character encoding was mangled and the bytes for "1" and "5" should perhaps be in a crazy encoding (think ebcdic on steroids) that maps them to "7" and "-2".

So I guess assuming the date "1/5/10" could refer to the year 2BC is about as logical as assuming it could refer to 2001?

And the 10th of May, 2001, wasn't 10 years ago anyway (not that it is relevant)...

My all-time favourite was .NET running in a Thai locale interpreting standard "completely unambiguous" ISO yyyy-MM-dd as a Buddhist calendar year, i.e. Gregorian year + 543. Hilarity ensued.
I don't see where 10ya comes from, but DMY, MDY, and YMD, at least, are all plausible interpretations.
YMD is only plausible if the first number is 2-digits or 4-digits. How is "1" a year? Unless we're talking about 1AD...
> YMD is only plausible if the first number is 2-digits or 4-digits. How is "1" a year?

During the first decade of the 21st Century, I regularly encountered dates, especially handwritten, with single digit years, because some people first reduce to a two-digit year when writing dates, but then -- as with any other number -- leave off leading zeros.

Really? I don't think I've ever seen this, though I guess I don't deal with a lot of handwritten dates.
The date in question ("1/5/10") was formatted by a computer, so hand-written dates seem pretty far off topic.

But even ignoring that, I've never ever seen a single person write a single-digit year. Not alone ("I graduated in 1" vs "I graduated in 2001"), not in dates ("August 7, 1" vs "August 7, 2001"), not anywhere.

The single digits in the date are not prefixed with a zero. So "1" could be interpreted as 2001. Why would the first digit be prefixed with zero but not the middle digit ("5").
One of the comments says:

> when I searched on the problem just now, I found complaints about the same problem going back to 2006.

(comment deleted)
(comment deleted)
Using a British English copy of android in Firefox the date of the first post is 05/01/2010 for me...
Yeah, with my Japanese locale it shows up as 2010/01/06
Perhaps it's 2006 years old?
How can this not get fixed after 10 years? You'd have to assume that googlers use their own Calendar, so why aren't they out with pitchforks to incentivize the programmers? Do the in-house Calendar feeds get refreshed every hour or so?
Sometimes when I use a Google product other than search I wonder if it's more cynical to believe it's in the state it's in because they don't dogfood it, or because they do but they still don't care.
It's probably a real pain in the ass to fix - with some dumb tech debt reasons why.

But, what's probably contributing the most to it is how Google incentivizes people internally. The promotion process is tied very heavily towards launching new features.

Maintaining existing products is (often) a bit of a career dead end for Googlers. It's safer to chase some new and shiny feature.

FWIW: Facebook's events feed syncs quite fast in Google Calendar. Certainly a lot faster than feeds of less known sources.

My naive assumption is that they prioritize by popularity?

(comment deleted)
Reminds me of what's going on right now in Chrome:

https://bugs.chromium.org/p/chromium/issues/detail?id=608016...

For all of us HN readers, backspace to go back made intuitive sense from the first time we accidentally hit backspace years ago. However, I imagine for every one of us, there are tens of folks who use computers sparingly and get confused when hitting backspace takes them to a wholly different page. In the end, this might make chrome easier to use for the common man at the expense of a tiny performance hit for power users.
I'm a user of Firefox, and have backspace to go back disabled, due to having accidental navigation happen way too many times. When I do want to go back, I can always press alt+left
Problem is that you can't press alt+left with one hand. Backspace wasn't reliable but at least it worked when you tabbed away from any active inputs.
You can use alt+left with one hand, just use the right hand alt key.
Many keyboard layouts have altgr instead of a right alt key, and it won't work for this.
This is what user-specific settings are for.
Each user-specific setting introduces additional maintenance overhead and rarely used/dead code. Easy to see why they'd want to avoid it.
I have to disagree. I've not dug into customizing Chrome, but creating a mapping for key -> action is pretty much bog standard GUI work, and doesn't require rarely used/dead code. Key presses are captured by the GUI, which maps them to an action. If users want to keep backspace as back, that should be just as easy as letting someone make vi keybindings for navigation.
The Chromium team has a policy about these things and most of them are larger pieces of code (this one is an exception).

Apparently this case isn't special enough to make an exception, which is understandable since you can just use an addon to restore the old behaviour.

I just use Vimium, 'H' is the default shortcut to go back a page. Mind you, Vimium also supports custom hotkeys, so that would be a possible way to get that back.

And it's really good for full keyboard navigation. What I'm trying to get at is that, power users have options.

To me, personally, backspace made sense but even as a power user of sorts I'd say I've used it accidentally more often than not.

Except what they went with is the bad and lazy solution.

Other browsers have already implemented the proper solution, which is to cache form inputs across history navigation, but they simply didn't care enough to actually improve the user experience.

i first encountered this bug^H^H^Hfeature yesterday. thought i accidentally enabled or changed some obscure setting or hotkey.
Definitely the first thing I thought of as well. My experience with this change has been so negative I've actually been seriously considering going back to Firefox, all these years later.
Firefox also doesn't support backspace to navigate backwards.
Sure, but my guess is they will either let me remap or have an extension that doesn't look at and control all my browsing data to do it. I haven't looked yet, though, so I might be wrong (I just remember having more flexibility when I used Firefox).
Firefox is currently in the process of changing their extension API to more closely match Chrome's
(comment deleted)
I've lost form content so many times from backspace going back a page (thumb just glances the touchpad to change focus from editable control).

Backspace isn't a navigation function. It's to delete the last character. We have navigation keys they're called cursors.

Glad to see the back of backspace.

Personally, the amount of time I've spent re-typing data into a browser form because of accidentally pressing the backspace key while a field wasn't focused (usually caused by brushing the touchpad on my laptop) far outweighs the speed benefits I gained from using backspace to navigate back into my browser history over the past 15 years.

However, it would have been nicer for the Chromium developers to allow users to remap key bindings instead of removing the binding and telling users to use an unpublished extension to add the functionality back.

Imagine the outcry if all of a sudden Microsoft decided that the backspace key in Word should no longer work, and suggested that users who wanted the functionality fixed should go download and install a third party plugin.
Backspace is totally useless. Not sure about Chrome, but in other browsers you can use alt+left arrow alt+right arrow, which is more intuitive and harder to do accidentally.
Hotkey use is a very habitual and personal thing. Everyone has a scheme that works for them and it's hard to adapt to something else. So why don't modern browsers just add a screen where you can customize your hotkeys?

Like, you know, the entire video game industry has managed to do for decades?

Different use case with different concerns.

For starters, you don't expect to be able to sit down in a public library and have their freely-provided copy of Elite: Dangerous work the way it does at home.

Just learn to use ALT+LEFT ARROW (back) and ALT+RIGHT ARROW (forward).

Backspace was problematic because it was too easy to accidentally activate, plus you'd still need to know about ALT+RIGHT ARROW to go forward again which wasn't consistent.

They were right to remove it. Terrible hotkey.

> plus you'd still need to know about ALT+RIGHT ARROW to go forward again which wasn't consistent.

Or, you know, use shift+backspace, just like you can use shift+space to scroll back up.

I think that's a different issue. The Chromium change was very intentional to work around an otherwise-unavoidable failure mode users would encounter (some pages don't restore correctly if you navigate off of them, and backspace also meaning "delete previous character" meant that having it overloaded as the "go back a page" operator too often led to undesired navigation events).

Incidentally, aren't there already Chrome extensions to change the behavior to what it was originally? [https://chrome.google.com/webstore/detail/backspace-as-backf...]

use this oss extension and keep going... https://github.com/j-delaney/back-to-backspace

I'm happy backspace is gone. It's non standard UX behaviour in comparison to other browsers like Safari & Firefox etc.

Better use alt+arrow keys (this works also in all other browsers).

Google has annoyed me in the past with Chrome feature removals (I want my fast profile switcher back!), but this one I agree with.

It's easy to see how people would loose data by accidentally using backspace. I'm fine with Alt-Left and Right click -> Back.

The right click -> back method sounds clunky, but it's actually extremely convenient.

Have you tried Vivaldi? Chromium-based, very configurable (including keys), lots of features, and they are always interested in adding new ones.

I dropped Chrome after Vivaldi came out, but there are plenty of other Chromium-based browsers to try....

Well the most important reason to use Chrome is its unrivalled security track record. Getting it right takes skill, resources and time, and Google is hard to match.
You get what you pay for.

My calendar on my phone routinely forgets to remind me, even though clearly the entry says "remind me 1 hour before" and other events with the same reminders work fine.

Are there any good calendar hosting companies where I could pay a few $$ a month for decent features, bug fixes, and service?

5 minutes ago my FastMail calendar sent me an email reminding me of a dental appointment tomorrow morning. FastMail in general seems reliable (except for the long outage they had 3 years ago).
Thanks, I'll look in to it. Hopefully reminders come in forms other than email, there's a decent android app, and I can access an exported ical format programatically for backup purposes.
Fastmail is quite reliable, and they support CalDAV and CardDAV. Their android email client is fast but doesn't work offline--fortunately they support IMAP Push and work very well with whatever 3rd party email client you prefer. I haven't moved my personal domain over there yet, but I've been using a paid tier for several emails with them for over 10 years and have been very happy. I haven't tried the exported ical format programmatically, but some of their devs are active on HN and if you shoot their help desk a question about it I expect they'll get back to you promptly.
My iOS calendar, synched with Google calendar, always works. Been depending on that form of reminders for at least two years now, and I can't recall ever being let down by it.
Thanks. I run Android; maybe I just need a different app to get the reminders right, but still sync to Google's calendar, since it's free and available everywhere.

Or maybe it's time to brush up on Kotlin and just write it myself.

I was wondering the same thing yesterday.

Searching around only returned me smartphone apps that connect to Google Calendar, with a couple of actual alternatives (Zoho, which I'm using as a primary email provider and can't set the freaking timezone right, and Outlook calendar, where I absolutely despise its web design).

It's like nobody's even trying to implement this properly. I'm trying out Yahoo's calendar now.

(comment deleted)
My iOS calendar, synched with Google calendar, always works. Been depending on that form of reminders for at least two years now, and I can't recall ever being let down by it.
Take a look at posteo.de

A German privacy focused mail provider where you also get a calendar. I didn't use the calendar yet but their mail servers are faster than anything else I've seen. And you can encrypt your whole account on their server so that noone but you can read your mails.

Edit: their English website is https://posteo.de/en

I've been meaning to turn on birthday reminders on Google Calendar for a while now. It turns out it doesn't even have such a feature.
Partially incorrect.

While Google's calendar does have a "Birthdays" calendar, it only includes birthdays from your Google+ contacts, and there's absolutely no way of adding a custom birthday into it. So, you would have to have a Google+ account and add every single person you want to include birthdays in as your Google+ contact.

Outlook's calendar has an option of connecting to Facebook and collecting birthday info from your friend list if you wish to. I know that Russian Mail.ru calendar also has this option (+ optional email alerts for birthdays). Pretty handy if you keep your Facebook friend list nice and short. You don't have to add all of them manually.

Important question: Are you using Google's Calendar app or are you using the OEM calendar app that came with your phone?

It's an important distinction. The second rate calendar apps pushed out by the phone manufacturer are generally pretty buggy and hardly ever see proper updates for security or bugs.

Solution: Download Google Calendar from the Play Store here: https://play.google.com/store/apps/details?id=com.google.and...

Open it up and give it access to your calendar, contacts, etc as requested. Then go and disable the calendar app that came with your phone.

Seconded on Fastmail. They also don't seem to be interested in selling your info.
> You get what you pay for.

My company pays quite a bit for Google services (including calendar), and I personally have had this problem bite me for meetings which don't show up in the feed until well after they've occurred.

Does your company get support for its money? If so, ask them to increase the priority of fixing the bugs you are seeing, since you are paying them.
This is how it would work at a company that's not completely deaf to user complaints

Unfortunately this is not the case of Google

This seems to be the case for users who don't pay for Google services, but all of the companies I've dealt with who pay for Google services seem to get responses to their problems.

If you are paying for Google services and you aren't getting your money's worth, look elsewhere, IMO.

It's the cult of the API.

I have personal experience with this. Because of the lack of reliable iCal sync, a project I worked on had to spend over a week and a half of developer hours implementing, troubleshooting, and maintaining Google Calendar API integration for our users. It's the source of no shortage of pain for this app as it is a sideline feature, but has just enough demand that Google integration cannot be dropped.

Google doesn't want to support iCal because they want to have everyone use their API.

Same for IMAP being deprecated in favor of the Gmail API, which gives access to many more features.
For clarification:

What does the Gmail API add over IMAP?

Google doesn't care that the Picassa system's servers are used for bouncing spam. The spam arrives with a from-envelope address of "<random-id>@photos-server.bounces.google.com". This has been going on for years and reported to Google repeatedly. They don't give a damn.

I bet you my e-mail server has caught this recently. Let's see:

  # zgrep photos /var/log/exim4/rejectlog*
  /var/log/exim4/rejectlog.17.gz:2016-07-24 04:34:41 H=mail-qk0-f201.google.com [209.85.220.201] F=<3y6eUVwYUADANRSTQeNRSTQe.kkPQeeM.ebMOQ@photos-server.bounces.google.com> rejected RCPT <ada-mp1-request@kylheku.com>: Too many components in domain name
  [... snip, a number of hits going back to May ... ]
There we go: the most recent spam attempt was on July 24, 2016. Someone tried to use Picassa to spam my ADA MP-1 mailing list.

So that's how I'm catching this stuff now: my rule against too many components in the sender domain is taking care of it, currently, before anything more specific.

Here is one 2012-dated discussion about this:

https://productforums.google.com/forum/#!topic/gmail/bQd_kZl...

It goes back before that, though, to at least 2010.

"I hope this can be resolved soon..."- dawnjc79 4/17/12

poor guy

I think Google cares, but they're not communicating it very well. I would hesitate before building a feature that might kick off a relatively complicated process with a lot of external factors whenever a user wants. It seems very complicated and without much positive benefit for the average user.
(comment deleted)
hey, other google users which are volunteers are answering all your questions. You are not supposed to talk directly to Google! ;)
So, much fun as it is to bash Google, this is a feature request. They're not obligated to implement a free enhancement on any sort of timetable.

This isn't like when Google actually broke the telephony in the Nexus 4 and then finally closed the issue with a middle finger: https://code.google.com/p/android/issues/detail?id=82949

My mind was blown when I discovered this little gem recently:

http://www.computerhope.com/unix/ucalande.htm

Place a text file at ~/.calendar/calendar with one event per row, containing the date and the title separated by a tab. Type calendar and you'll get a list of what's coming up in the next few days. You're done.

Being a venerable UNIX utility there are of course a variety of options (switch to view X days in the future, support for including external calendars, support for recurrence, etc.). Weirdly, it lacks any concept of start and end time, but this is easy enough to simply include in your appointment's title. And no free/busy or special sharing mechanism (though you could certainly compose your calendar of several files and expose them on a web server easily enough).

It's not about to replace a full-blown calendar app for power users, but if you're the type of person who has a fairly light meeting schedule and spends a lot of time in a terminal... I was immediately done with all other calendar software and consequently all the problems of said software when I stumbled across this thing.

As someone who builds a scheduling product which offers an iCalendar feed for Google Calendar (among others), I have to face this issue on a weekly basis.

The most plausible explanation I came up with (and what I tell my users) is that Calendar is maintained by a small team in Switzerland for which this isn't a priority. I don't blame anyone, if true. It might not be easy to fix with little resources.

The fact that some feeds sync faster than others (Facebook events) and syncs speed up as time passes, makes for a weird case.

Outlook.com or iCloud/iCal sync properly. One of those offer a way out for my users if critical.

Google is all about scale, and caring doesn't scale. This is part of their DNA.

This might seem like a flippant comment, but it explains a lot of the observations I've made of Google over the years.

They'd have a lot more clout if they were customers.
Was this hidden from the front page?
Is this just a problem when subscribing to calendars published through Google? In other words, if I host an iCal calendar on my own server, will subscribers see changes immediately when they refresh their Google Calendar client?