Auto save should have come before those advanced features...

I have a mouse in my desk but almost never use it.

I guess I'm good at hotkeys without really realziing it.

Auto save on visual studio is lololol awful. It's the first thing I change. It renders the whole computer unusable every 5 mins.

For vs code, save on focus change is nice tho.

At some point, Excel takes a long time to save anything. Maybe for a basic 20x20 CSV file, but once you have 100x1000, VBA, functions, etc, a save can be a precarious operation in itself. Doing that automatically while trying to type other stuff is asking for many more crashes, freezes, and corrupted files than you get with normal use.

Then make it an option.

Another MS bugaboo. When I change a meeting in outlook from 11am to 1 - just go ahead and change AM to PM for me you ****sticks. Can't tell you how many meetings I've screwed up by setting them for 3am. Arghhhh. It's like the old days of setting your VCR to program when several consecutive miracles had to happen for it to actually work.

I can't even.

Outlook doesn't mark an email read until you deselect it, which just drives my OCD crazy seeing the new email notification in the task bar.

This bugs me too. AFAIK it's been this way for like... decades. I usually either delete a message once I've read it, or select away and back. I'm one of these guys who is always at zero unread messages.

This used to bother me too until I realized this is an option that you can change.

ETA: File -> Options -> Advanced -> "Reading Pane..." button -> click stuff (I forget what it looked like before, but I have "Mark items as read when viewed in the Reading Pane" checked and "Mark item as read when selection changes" unchecked.

Then it becomes a support nightmare - you can't have a user setting that can break lots of third-party software for instance. Every new version of Excel had to compete with an old version of Excel - it's hard to win at the enterprise level if you're breaking existing integration. Also certain antivirus software will freeze Excel a couple of seconds every time you save, if you have a sufficiently complex workbook with VBA. Lots of traditional enterprise shops also use slow and unreliable shared drives. USB sticks or other slow and/or unreliable removable media pose an issue. Maintaining a platform like this is like maintaining a complex legacy code base, except you don't even have access to most of the code base. You have to keep all of that working without having any idea what it is, let alone the ability to change it.

In general, it's quite difficult to change 20+ year old software with a massive ecosystem that's built around it in some radical way to streamline the workflow for a few users. Excel is a platform, not an application.

This almost certainly is due to poorly written VBA, add-ins, third-party anti-virus, etc.

Come on dude. Every IDE can auto-save. Excel could do it too. Just gimme the option, doesn't have to be default.

It's a freaking macro - if (timeElapsed(> 1 min) then save. onLoseFocus(save())

You sound like Kerowo when it comes to Apple

Hah, yeah its really not a big deal to implement auto-save. There are lots of harder problems that have been solved.

That's not touching the argument of whether it should be a new default or anything. Arguments for and against that.

A common operation I had to do was take a 250 x 3000* book with blank rows every 10 items or so. I then had to open a new column and run a macro for copy / paste / move. Once complete, remove all blank rows and save. It's definitely a precarious operation. That's the problem with auto-saving. If you are running a macro, how exactly would they build Excel to save while data is changing that fast?

But you are correct, some (all?) plugins are awful to use and can really cause havoc on a system.

* technically, I had to do this for 300,000 rows, but Excel wouldn't open at all with that amount of data. I think the upper limit is about 2M cells in a standard book, assuming there is little to no sparse data. A macro-enabled book or a book with a lot of blanks is considerably less.

Whenyourerightyoureright.jpg

That's the point though - Microsoft obvious can implement this if they want to, but this breaks enough things that they decided not to. Lots of Excel add-ins attach handlers to this event for instance:

https://msdn.microsoft.com/en-us/vba...ve-event-excel

Do you trigger this on auto-save? In some cases, you would want to, in some cases you wouldn't want to, but there's no good way for Excel to figure out which ones are which. Excel, through COM, can also be run in a headless mode through a scripting host - should it ignore these settings or not?

Also, if this was important to people, they could trivially download an add-in that does this. For instance Microsoft shipped an optional add-in that does this exact thing in Excel 2000 before they had the current auto-recovery feature. It just isn't that useful - Excel practically never crashes for non-power-users who don't use VBA stuff (for whom compatibility concerns like the above do matter) and the current autorecovery works better for crashes than overwriting the file you're working on, which is a huge no-no in some industries.

There are a billion add-ins out there and there's no shortage of "why doesn't Excel just do X" type of features.

What is much more useful (though technically close to impossible given the current architecture) is restoring the exact workplace setting, which is what IDEs, browsers and other applications are aspiring towards.

Somehow I knew I had a few more paragraphs on this issue in my future.

MS went through all the hassle of adding auto-save, they could easily implement it more than one way. If I can push CTRL-S at any time, they can auto-push it for me every minute. No I am not going down a rabbit hole on this.

Suzzer, to copy / past my question about that:

If you are running a macro, how exactly would they build Excel to save while data is changing that fast?

When you do what I described above, it's already edged up against full-on crash, reboot the computer, pray you aren't losing a week of work.

Software at this scale is more like a government agency than an appliance, especially since in many cases, there are no acceptable alternatives due to the network effect rendering the competition moot. Every product has to make trade-offs since they can't ship every feature and some features and workflows conflict with one another. You can't please everyone. This is normally okay because people have choices and have to live with whatever they choose. But when there is no choice, people feel trapped and unhappy with even relatively minor inconveniences. Because there's no way around it and the choice was made by others.

This is sort of the genius of Apple from a branding perspective btw - Apple, even at its best, is only a little bit better than their competition so people get to choose Apple products and feel empowered and invested in its success as to justify their conscious choice. Facebook, Microsoft and Google, in their areas of strength, are/were so much better than their competition that there is no realistic choice to be made. So the customers feel like they have no choice and they aren't as invested in the products and get to focus on the negatives.

This is actually not hard - all of VBA runs in a single-threaded event loop. They can have the autosave run in the same event loop. They already have to account for this to do autorecovery properly.

The real issue is that autorecovery as presently implemented is already a better version of the autosave that is being asked for here, for a number of reasons I already gave. So much so that the latter originally existed and was replaced by the former. Also, once you have the autorecovery feature, it would be highly awkward from a product perspective to then also offer a slightly different version of that feature.

Then don't use that auto-save option if you use a lot of macros. Not everyone uses Excel for ridiculously heavy macros. Why should that be the lowest common denominator?

Or something like if !macro.isRunning() then save(). The point is it's not really a macro, it's system level. Just auto-push the damn save button for me unless something is going on where that would be bad.

I'm not really a big fan of the "gosh I've thought about this for a full 30 seconds and can't figure out how it could be done, therefore impossible" line.

The bolded is just a bizarre assertion. MS offers 80 million different versions of different features all over their ecosystem. Go look at the Outlook options for marking things read.



Btw the reason this is a problem for me is I use an Excel spreadsheet as my todo list, organizer and dumping ground for any kind of notes. So I have it sitting open on my desktop at all times. Maybe I'm rare in that regard, I dunno.

I guess I take issue with the connection between using an IDE and using Excel. If you watch an expert programmer work, the pace is fairly leisurely. If you watch a professional Excel user work, you wouldn't be able to follow what he or she is doing.

The lowest common denominator of someone that uses Excel every day on their job is very different than the casual user of Excel. The difference between a casual and professional programmer (wpm, pacing, etc) is fairly close.

For example, it wouldn't be uncommon for someone using Excel to do some equation that more or less find / replaces 10,000 values. This isn't something you do every day in an IDE (I bet it crashes and freezes if you try).

Sublime can handle it, might take a minute or so.

This was a bad example but the lowest common denominator just doesn't matter here though. Excel is an enterprise product where purchasing decisions and network effects are almost entirely driven by advanced users. This isn't Word or PowerPoint, which are closer to mainstream consumer products.

Options for the same feature are distinct from two different features doing a similar, but different thing though. Also, there are no consequences for this feature and third-party products are unaffected by this choice. Offering "autosave" and "autorecovery" which do completely different things in a way that support has to account for and oh btw, if you use "autosave" some add-ins stop working is a non-starter. And because autosave/autorecovery are only useful in unusual, unanticipated circumstances, it's unreasonable to expect the user to choose the right option - no one cares and the default will prevail, except in circumstances where the user clicks random options to troubleshoot something else, which just adds to support burden. The only reason anyone has two deep features that are this similar to one another is if they start with one and want to switch to the other but can't abandon the previous one due to compatibility reasons.

I think I mentioned this here before but this is a classic example of how better software products necessarily have more detractors. Some software products are so good that they end up becoming successful well outside of its niche, at which point they end up being criticized for not being perfectly suited for something they weren't really intended to do.