I've personally looked into this some. Of the things on that list, SparkleShare is the only thing I've considered; I personally find that "third-party things tend to break" is not nearly as large a concern for me as "I might spend all day trying to get this to work and either fail or break it tomorrow." I also find that things designed to be accessible to a nontechnical audience tend to appeal to my sense of aesthetics much more, and to address more of my personal needs and concerns, as my use case scenarios are much closer to those of a nontechnical audience.
For me the biggest con of SparkleShare is that everything of mine would be made public, since I don't have my own server capable of running git (I use shared hosting). Another big problem is that I use an Android smartphone and a Nook Color as well, and would like access to my files on those devices.
Because it works with all of the above, I'm currently using Dropbox for sync, and I have a lot of things on other third-party services like Evernote as well. I'm considering moving to OwnCloud hosted on PortKnox though, which is already where my calendar is, and I'm using OpenPhoto for image hosting and a self-hosted WordPress install for writing (although I need to finish setting it up). Using GNOME's web browser to set up my WordPress site's admin console as a web app was exquisite; the distraction-free writing mode really is.
Edit: I'm also considering using my SparkleShare folder as a way to manage the open-source projects I work on, since it would sync with Gitorious without my having to intervene. All I'd have to do is set up the project online, and then change the directory I save things in ... if I understand correctly.
Edit x2: I also find that tools designed for a nontechnical audience tend to require less cognitive overhead to use. It's different if I'm already familiar with something and tied into using it for nontechnical reasons, which is one reason I write on Dreamwidth much more than on WordPress right now. But for me this usually makes the difference between "I can do this but I don't", and "I can do this and I enjoy doing so."
Edited (Added a couple of additional paragraphs) 2012-09-16 22:12 (UTC)
Of the things on that list, SparkleShare is the only thing I've considered; I personally find that "third-party things tend to break" is not nearly as large a concern for me as "I might spend all day trying to get this to work and either fail or break it tomorrow."
I was using SparkleShare for quite a while for that exact reason: ease of use. Unfortunately, as I mentioned in my article, the latest update (on Arch Linux) for whatever reason, has become impossible to use; it has basically broken.
And my concern isn't only that "third party things tend to break" but that "third party things tend to vanish". FLOSS solutions are less likely to vanish, because there's the possibility of someone else taking over the code base if the original author is no longer willing or able to continue with it.
As far as ease-of-use goes, I'm finding that gitdocs is sufficiently simple; you simply add or create the directories you want it to monitor. If one prefers using a GUI, one can use the web interface.
since I don't have my own server capable of running git
Can you use ssh to connect to the server? Because the only requirements for a "git server" is (a) ssh and (b) git itself. Are the shared hosting admins open to allowing git to be installed?
I also find that tools designed for a nontechnical audience tend to require less cognitive overhead to use.
Oh, certainly. That's one reason I switched to Ubuntu for quite a while, because it was a distro that required less effort to administer.
What I find, though, is that there's a trade-off; things that are simple to use are simple to use because they have fewer options to worry about. And sometimes I do want to go to the greater effort in order to have more control and more ability to customize the thing to do exactly what I want. For some things, I choose ease-of-use, and for other things I choose customization.
I'm also considering using my SparkleShare folder as a way to manage the open-source projects I work on, since it would sync with Gitorious without my having to intervene. All I'd have to do is set up the project online, and then change the directory I save things in ... if I understand correctly.
Yes, that would work, but there are limitations. - Sparkleshare doesn't seem to understand branches. - Sparkleshare always rebases the commits. - You wouldn't be able to write your own commit messages.
That's why I haven't included my open source projects in the list of git repos that I auto-sync, since most of the other auto-sync solutions have similar limitations. Again a case of simplicity versus control.
I haven't actually tried SparkleShare yet, but I've personally lost hours and hours and hours to fussing with git and I've never needed to do anything with it beyond committing my work. Switching from my current setup (saving everything in a "Projects" folder) to a setup where the folder I save everything to is shared and synced and versioned is an improvement in every way, if the projects are meant to be shared and aren't already part of another VCS infrastructure.
I personally feel that longevity is a quality a service can have independent of being "third-party" or open-source, and that it can be measured on its own. Simply being open-source doesn't give a project traction, and in my experience is not a meaningful factor on its own unless there is also a decent (both in size and behaviour) community committed to it.
I also feel that it needs to be balanced against the cost of switching. For instance, if Dropbox goes down, I'll just switch to another commodity service like SpiderOak; if Pulse goes down or puts ads everywhere, I'll go back to Google Reader. Conversely, I won't even look at Facebook because of how much it would cost to extricate myself, and I'm keeping an eye out for an Evernote alternative because I'm tired of switching notetaking solutions, even though Evernote is delightful and convenient and seems to have a good business model. I've tried SimpleNote and Tomboy, but both are unavailable or inconvenient to use across all my devices, even if they use more portable formats and / or are open-source.
The name "gitdocs" turns me off to the project as soon as I see it, for roughly the same reason seeing a GUI app prefixed with "X" does. It's a cultural marker that tells me it was made by and for people with very different needs and wants and abilities than me, for whom git is a beloved tool rather than an endless source of frustration. I'm open to being proven wrong, but I've had a lot of bad experiences with the free software community and its insular culture, and I strongly prefer using tools that were designed to be accessible to a wider audience than just them.
I don't have any particular desire to run my own instance of git, and if I wanted to roll my own backup solution I'd use OwnCloud. I don't personally feel confident in my ability to install outside PHP scripts aside from WordPress, though, because the few times I've tried it's gone horribly.
I feel that WordPress strikes a good balance between ease-of-use and customization, in that it's customizable to an amazing degree but designed to be simple to use for its core functions. I don't feel it's good to burden people with cognitive overhead in the name of "customizability," when most of them don't want to worry about how the thing they want to do happens.
I don't feel it's good to burden people with cognitive overhead in the name of "customizability," when most of them don't want to worry about how the thing they want to do happens.
You say that as if the programmers are forcing people to choose non-simple tools.
Yes, there do exist systems which have managed to get themselves the happy medium of simple-on-the-surface-but-customizable. But usually it's a choice between one and the other; that the tools that are simple are deliberately made non-customizable in order to have that simplicity, and the tools that are extremely customizable necessarily have added complexity in order to offer that customizability.
Take my favourite window manager, Fvwm. It does not offer a GUI for configuration, because any GUI would either have to be incredibly complex, or cut out half the options one can have for the settings. I love this window manager, because you can make it do practically anything. But the cost of that is that the configuration is done with text files, not a GUI. Many people may not want to pay that cost, sure. But I don't want to be deprived of my lovely window manager because some people find it too hard.
I don't like the idea of forcing someone to use tools that don't fit their lifestyle and workflow. But I feel that that's what many free software communities are all about, is privileging the kinds of people whose lifestyles and workflows fit existing tools and insulting / marginalizing people whose don't. I feel that a lot of the outcry about things like GNOME and Fedora's changes is that of privileged people having their privilege taken away from them, and having to learn a handful of new things so that many, many people don't have to make their brains fit into places they can't.
I'd rather not inconvenience anyone if possible, but my sympathies are with those who are currently being left out. I feel that there is a great deal of unspoken contempt for nontechnical users in many free software communities, and the ones that I have the most respect for are the ones that put the most effort into reducing cognitive overhead, technical / cultural knowledge requirements, and other obstacles to accessibility. It disgusts me to see things like white cismale hackers plastering the Fedora logo over their hackergotchis' mouths in protest against the idea of Girl Scouts on their system.
Indeed, the dark underbelly of FLOSS is truly ugly and horrible. I don't know the incident to which you are referring, but I am alas not surprised. There are only two FLOSS projects I am aware of that are female-welcoming, and that's Archive-of-our-own and Dreamwidth. The others at best are neutral.
I don't have much respect for projects that are deliberately elitist and newbie-hostile. Even though I'm not a newbie, they are unpleasant places to be, and I have no desire to contribute to them.
GNOME has actually been a lot of fun to work with. They run an outreach program for women that's explicitly trans-inclusive, and that actually pays women to work on the project. I spent my summer working on GNOME's JavaScript developer docs, and am hoping to mentor someone in the next round. GNOME's executive director is a woman, and I think the last one was as well.
I know GNOME isn't your preferred desktop, but I personally love it and it's also the one I'd feel most comfortable giving to a newbie. I feel that free software like it, which puts a high priority on being accessible to everyone, is very important to have around.
I don't agree with your dichotomy between simplicity and customizability.
[Error: Irreparable invalid markup ('<user=jewelfox>') in entry. Owner must fix manually. Raw contents below.]
I don't agree with your dichotomy between simplicity and customizability. <user=jewelfox> provides a good example with GNOME, and I'd even include things like Kate/Gedit and the not-FLOSS SublimeText.
All of these items are, in fact, very simple to just pick up and start using. A person can figure out the basics of what they want to do within a very short period of time, and be happy with that.
Then, the power users can dig around under the hood and customize them to their heart's content. Kate and Gedit have their plugin systems that allow Gedit to go from a step above Notepad to an editor that rivals Mac's TextMate. SublimeText has the plugin system, plus a simple-to-understand text-based setting system (ST is, of course, aimed at developers, so it's expected that such users wouldn't be intimidated by config files, and most of it's simple key-value pairs). Gnome has extensions, plus the ability to edit the Javascript and CSS files that control just about everything (which has allowed for variations such as Cinnamon and Unity, as well as themes in general).
A lot of basic usability/simplicity (and Jewel's desire for lower cognitive overhead) isn't so much in the lack of customizability, but rather a fantastic set of defaults, so that it doesn't <em>need</em> to be customized. For example, as much as I love toying around with GNOME's CSS file and themeing stuff, I had no desire to do so on my Fedora install, because I already liked the defaults. I think I actually did more customizing of my Ubuntu 12.04 (with Unity) install than I did with my Fedora one (and Unity has been notoriously hostile to customizability until the most recent version or two; and even with Ubuntu Tweak, there's still quite a bit that Canonical has locked out completely).
Look, I don't even understand why we're having this argument! All I wanted to do was share my findings about DropBox alternatives, not argue about user interfaces!
Don't get me wrong, the fact that you shared what you found is awesome, and I'm sorry that that was glossed over.
However, part of what makes an application appealing, regardless of whether it's FLOSS or commercial, is not only finding that balance between customizability and ease of use, but providing good defaults. It seems to me that your assumptions are that "easy to use" means it can't be customizable, I think your choices in your original post (as well as your comments here) reflect that, and I commented accordingly.
Also, I wasn't arguing, but rather joining in the discussion, which I felt was interesting (though perhaps that was missed, given that it appears I screwed up the DW/specific markup and it butchered the whole rest of my comment?) and worthwhile in the broader topic of tool choice.
Am I the only one that sees irony in the "third party things tend to vanish" line (considering the commerical world basically says "use commerical stuff, because FLOSS things tend to vanish")?
On a side note, regarding git-based solutions in general - depending on what, exactly, one is looking to accomplish with these solutions, another option is to pair git with Capistrano and a local file-change-watching daemon to automate most of your syncing. The trade-off here is, of course, no unified GUI interface, but tons of control.
no subject
For me the biggest con of SparkleShare is that everything of mine would be made public, since I don't have my own server capable of running git (I use shared hosting). Another big problem is that I use an Android smartphone and a Nook Color as well, and would like access to my files on those devices.
Because it works with all of the above, I'm currently using Dropbox for sync, and I have a lot of things on other third-party services like Evernote as well. I'm considering moving to OwnCloud hosted on PortKnox though, which is already where my calendar is, and I'm using OpenPhoto for image hosting and a self-hosted WordPress install for writing (although I need to finish setting it up). Using GNOME's web browser to set up my WordPress site's admin console as a web app was exquisite; the distraction-free writing mode really is.
Edit: I'm also considering using my SparkleShare folder as a way to manage the open-source projects I work on, since it would sync with Gitorious without my having to intervene. All I'd have to do is set up the project online, and then change the directory I save things in ... if I understand correctly.
Edit x2: I also find that tools designed for a nontechnical audience tend to require less cognitive overhead to use. It's different if I'm already familiar with something and tied into using it for nontechnical reasons, which is one reason I write on Dreamwidth much more than on WordPress right now. But for me this usually makes the difference between "I can do this but I don't", and "I can do this and I enjoy doing so."
no subject
I was using SparkleShare for quite a while for that exact reason: ease of use. Unfortunately, as I mentioned in my article, the latest update (on Arch Linux) for whatever reason, has become impossible to use; it has basically broken.
And my concern isn't only that "third party things tend to break" but that "third party things tend to vanish". FLOSS solutions are less likely to vanish, because there's the possibility of someone else taking over the code base if the original author is no longer willing or able to continue with it.
As far as ease-of-use goes, I'm finding that gitdocs is sufficiently simple; you simply add or create the directories you want it to monitor. If one prefers using a GUI, one can use the web interface.
since I don't have my own server capable of running git
Can you use ssh to connect to the server? Because the only requirements for a "git server" is (a) ssh and (b) git itself. Are the shared hosting admins open to allowing git to be installed?
I also find that tools designed for a nontechnical audience tend to require less cognitive overhead to use.
Oh, certainly. That's one reason I switched to Ubuntu for quite a while, because it was a distro that required less effort to administer.
What I find, though, is that there's a trade-off; things that are simple to use are simple to use because they have fewer options to worry about. And sometimes I do want to go to the greater effort in order to have more control and more ability to customize the thing to do exactly what I want. For some things, I choose ease-of-use, and for other things I choose customization.
I'm also considering using my SparkleShare folder as a way to manage the open-source projects I work on, since it would sync with Gitorious without my having to intervene. All I'd have to do is set up the project online, and then change the directory I save things in ... if I understand correctly.
Yes, that would work, but there are limitations.
- Sparkleshare doesn't seem to understand branches.
- Sparkleshare always rebases the commits.
- You wouldn't be able to write your own commit messages.
That's why I haven't included my open source projects in the list of git repos that I auto-sync, since most of the other auto-sync solutions have similar limitations. Again a case of simplicity versus control.
no subject
I personally feel that longevity is a quality a service can have independent of being "third-party" or open-source, and that it can be measured on its own. Simply being open-source doesn't give a project traction, and in my experience is not a meaningful factor on its own unless there is also a decent (both in size and behaviour) community committed to it.
I also feel that it needs to be balanced against the cost of switching. For instance, if Dropbox goes down, I'll just switch to another commodity service like SpiderOak; if Pulse goes down or puts ads everywhere, I'll go back to Google Reader. Conversely, I won't even look at Facebook because of how much it would cost to extricate myself, and I'm keeping an eye out for an Evernote alternative because I'm tired of switching notetaking solutions, even though Evernote is delightful and convenient and seems to have a good business model. I've tried SimpleNote and Tomboy, but both are unavailable or inconvenient to use across all my devices, even if they use more portable formats and / or are open-source.
The name "gitdocs" turns me off to the project as soon as I see it, for roughly the same reason seeing a GUI app prefixed with "X" does. It's a cultural marker that tells me it was made by and for people with very different needs and wants and abilities than me, for whom git is a beloved tool rather than an endless source of frustration. I'm open to being proven wrong, but I've had a lot of bad experiences with the free software community and its insular culture, and I strongly prefer using tools that were designed to be accessible to a wider audience than just them.
I don't have any particular desire to run my own instance of git, and if I wanted to roll my own backup solution I'd use OwnCloud. I don't personally feel confident in my ability to install outside PHP scripts aside from WordPress, though, because the few times I've tried it's gone horribly.
I feel that WordPress strikes a good balance between ease-of-use and customization, in that it's customizable to an amazing degree but designed to be simple to use for its core functions. I don't feel it's good to burden people with cognitive overhead in the name of "customizability," when most of them don't want to worry about how the thing they want to do happens.
no subject
You say that as if the programmers are forcing people to choose non-simple tools.
Yes, there do exist systems which have managed to get themselves the happy medium of simple-on-the-surface-but-customizable. But usually it's a choice between one and the other; that the tools that are simple are deliberately made non-customizable in order to have that simplicity, and the tools that are extremely customizable necessarily have added complexity in order to offer that customizability.
Take my favourite window manager, Fvwm. It does not offer a GUI for configuration, because any GUI would either have to be incredibly complex, or cut out half the options one can have for the settings. I love this window manager, because you can make it do practically anything. But the cost of that is that the configuration is done with text files, not a GUI.
Many people may not want to pay that cost, sure. But I don't want to be deprived of my lovely window manager because some people find it too hard.
no subject
I'd rather not inconvenience anyone if possible, but my sympathies are with those who are currently being left out. I feel that there is a great deal of unspoken contempt for nontechnical users in many free software communities, and the ones that I have the most respect for are the ones that put the most effort into reducing cognitive overhead, technical / cultural knowledge requirements, and other obstacles to accessibility. It disgusts me to see things like white cismale hackers plastering the Fedora logo over their hackergotchis' mouths in protest against the idea of Girl Scouts on their system.
no subject
I don't have much respect for projects that are deliberately elitist and newbie-hostile. Even though I'm not a newbie, they are unpleasant places to be, and I have no desire to contribute to them.
no subject
I know GNOME isn't your preferred desktop, but I personally love it and it's also the one I'd feel most comfortable giving to a newbie. I feel that free software like it, which puts a high priority on being accessible to everyone, is very important to have around.
no subject
All of these items are, in fact, very simple to just pick up and start using. A person can figure out the basics of what they want to do within a very short period of time, and be happy with that.
Then, the power users can dig around under the hood and customize them to their heart's content. Kate and Gedit have their plugin systems that allow Gedit to go from a step above Notepad to an editor that rivals Mac's TextMate. SublimeText has the plugin system, plus a simple-to-understand text-based setting system (ST is, of course, aimed at developers, so it's expected that such users wouldn't be intimidated by config files, and most of it's simple key-value pairs). Gnome has extensions, plus the ability to edit the Javascript and CSS files that control just about everything (which has allowed for variations such as Cinnamon and Unity, as well as themes in general).
A lot of basic usability/simplicity (and Jewel's desire for lower cognitive overhead) isn't so much in the lack of customizability, but rather a fantastic set of defaults, so that it doesn't <em>need</em> to be customized. For example, as much as I love toying around with GNOME's CSS file and themeing stuff, I had no desire to do so on my Fedora install, because I already liked the defaults. I think I actually did more customizing of my Ubuntu 12.04 (with Unity) install than I did with my Fedora one (and Unity has been notoriously hostile to customizability until the most recent version or two; and even with Ubuntu Tweak, there's still quite a bit that Canonical has locked out completely).
no subject
no subject
However, part of what makes an application appealing, regardless of whether it's FLOSS or commercial, is not only finding that balance between customizability and ease of use, but providing good defaults. It seems to me that your assumptions are that "easy to use" means it can't be customizable, I think your choices in your original post (as well as your comments here) reflect that, and I commented accordingly.
Also, I wasn't arguing, but rather joining in the discussion, which I felt was interesting (though perhaps that was missed, given that it appears I screwed up the DW/specific markup and it butchered the whole rest of my comment?) and worthwhile in the broader topic of tool choice.
no subject
On a side note, regarding git-based solutions in general - depending on what, exactly, one is looking to accomplish with these solutions, another option is to pair git with Capistrano and a local file-change-watching daemon to automate most of your syncing. The trade-off here is, of course, no unified GUI interface, but tons of control.