Topic: New bitcoin.org Clients page (Read 6082 times)

Glad you like it!
 
But I believe genjix is quite focused on the Bitcoinica issue at this time (which is fully understandable).

much better: http://dre.tx0.org/compare.htm (https://bitcointalksearch.org/topic/desktop-clients-quick-comparison-chart-83768)

link from bitcoin.org?

+1. Luke's table should be linked somewhere on the page.

is the table still available somewhere? the page as it is now is not in my taste.

maybe the comparison table should go into the wiki and a link to the comparison should be at the top of the bitcoin.org clients page.

I totally agree.
When someone wants to try out "this bitcoin thing they just heard about" they are not going to jump through a lot of hoops to get going. We have a few minutes (at best) to let them install and run the software. It is all about eliminating barriers of entry. Once a new user has tried out a "simple wallet" and had their first success with Bitcoin they will start looking for more advanced stuff.

However, as we have learned from MyBitcoin (and others), we cannot and should not place too much trust in centralized services. The temptation is simply too big when dealing with something that represents real value.
Fortunately the community has developed an impressive number of alternative clients that are swift to install and easy to use. Among those swift'n'easy wallets we should promote the ones where the user does not risk another MyBitcoin (god forbid).

Oh, did I mention that BitcoinSpinner for Android installs and is operational in less than a minute and keeps your private key on your device. It was designed with the above in mind from scratch. It does rely on a central server to send/receive coins, but you can always evacuate your private key independent of the server and import it in for instance MultiBit. (yes, I am the main dev of BitcoinSpinner)

Blockchain.info has similar offerings for Web/Android/iPhone and looks very promising.

+1

Simply put I think we should have:
1. a platform chooser: (Windows/Linux/Android/iPhone)
2. a simple question: Are you new to Bitcoin?

It's a fair point that even having multiple clients could be a bit overwhelming, but people read about Bitcoin in the news, they search for Bitcoin on their favourite search engine, it'll take them to bitcoin.org. If they can't find a good client there they'll just assume there are none.

If the Satoshi client was completely satisfactory to everyone it might not be a big deal, but for performance/features/other reasons it seems that multiple desktop clients are here to stay. Not to mention that Satoshis code would need a lot of work to run well on mobiles, so there'd need to be at least a page for them.

It's probably best to just accept the situation as non-ideal and muddle through as best we can.

Bitcoin.org is naturally "default" no matter what anyone does/says. Giving a few clients priority just because they were based on Satoshi's original code just makes Bitcoin more centralized. Bitcoin.org should either have a directory of clients, or none at all. Gavin mentioned the latter as his goal (comparing it to BitTorrent.org), but I think at this stage, people will get lost if they can't find anywhere to go for a client on Bitcoin.org...

If anything, this "reference project" should move off Bitcoin.org, to a site of its own, just like BitcoinTalk.

This page works, I can update it with mouse over descriptions and stuff.  It's quick and to the point.

http://lovebitcoins.org/getStarted.html

It will be getting ALOT of traffic very soon.

Personally, all this seems far too focused on a centralized website (http://bitcoin.org/), and presents far too many choices at once to the user.

On bitcoin.org (registered by Satoshi), I would rather see the Satoshi reference client and perhaps an "other clients" link on the wiki.

Modern websites are working hard to reduce the number of download links, not increase them.  See, e.g. http://fedoraproject.org/en/get-fedora or http://www.mozilla.org/ where a single download choice is presented, and then an "other options" link is below the great big download button.

You don't see people calling for Mozilla to link to other browsers.  Rather than fighting over what a particular bitcoin.org page should look like, why not maintain an independently managed BitcoinClients.org website?  Or GetBitcoinClient.org or somesuch.

Most of all, solve this problem in a distributed fashion, rather than stuffing it all onto bitcoin.org.  Bitcoin.org, IMO, is the home of the "reference project" not the entire bitcoin community.  Emphasizing that months ago was why the forum was moved to bitcointalk.org.

+1

And to add my 2 btcents to the topic, easy of install is really important. As much as I liked the concept beside Electrum, I was turned off immediately by the install page for linux... and never installed it.

Really, as a user, I want as few clicks and texts as possible. That's stupid, but we want to make bitcoins as user-riendly as possible.

Good point.  The average newbie starts with $5-$20, so for learning how wallets and addresses work, its perfect. 

+1

Only problem with this, is that they are not as secure.
That should be fine if it could be pointed out in a way that doesn´t sound to intimidating.

 

I can picture less computer savvy people being turned off by Luke's table because they wouldn't know what 1/2 the columns mean and whether they are important.  I personally prefer it, but I agree with the suggestion that it should be on a separate "compare clients" page, or on the same page but below the screenshot & summary descriptions.

People tend to have very short attention spans when they are searching for a download link.  It might be helpful to order the clients by ease of use, starting with the most general/friendly clients and progressing towards the advanced/resource intensive clients.  I think it is safe to assume that advanced users already know what they want, so there is no need to cater to them on a page that is intended for everyone.

What do you mean by "Electrum is not user friendly"? Are you serious?

Electrum is the only user friendly desktop client.

When i first opened Satoshi client, i thought i got a virus going on.

When i first opened Armory, my comp crashed.

When i first opened Electrum, i was ready to go in 5 seconds!!!


Regarding text... i believe 3 sentences are more than enough for brief info.

I know my opinion doesn't count much, but i had to put it out!

I'd prefer not to have the reviews independent as there is a potential for abuse there. As an example: say that one day I have a falling out with MultiBit (they are great guys btw ), and so I try to show them in a bad light. Such a situation is not good for anybody.

Best to find a situation which is amicable to everybody. I thought your descriptions were better written than mine for the average user, but ThomasV takes issue with some of the modifications. Fine, I can understand. Lets find those small differences and create something the antagonistic reviewer (me and you in this specific example) and the submitter (ThomasV) both agree on.

Original text:

Newer text:

Well I think we can agree that the opening line is more concise and much better:


Second sentences can both be merged perhaps:


The criticism is good but maybe a bit long, so lets include that but shorten it:

To follow up on that, I recognized how many of my own users were getting stuck at the installation process, so I finally dug into it a couple weekends ago and Got It Done, in Armory.  

I used Jgaa's "War Setup" for Windows.  It is pretty fast and easy after you get past a few hangups (like how to add your license if it's not on their default list).  I selected the "Minimal" install which literally just has an EULA and a "Install" button.  Done.  It's almost too easy -- the program is installed before you even realize that was the last button to press! 

For Linux, I made Debian packages to support Ubuntu/Debian, and that was actually a pain in the ass, but totally worth it once you figure it out.  Worth it because it's easy to tie it into your build process so that building the .deb package is just an extra couple keystrokes on the command-line build process.

I figured, anyone using a non-Ubuntu setup is likely familiar with the compile from source process, and Armory's isn't so bad.  So I left that as is.

If I ever get a Mac, I will make an installer for that too, but I don't know how difficult that is.

I've gotten tremendously positive response to making installers.  Hopefully what I wrote above can help others make them...

Perhaps the page is not clear about that, but there is not one install process, there is choice.
Linux users have the choice between install from source and a tarball that contains all the dependencies.
Windows users too have choice between a zip file and a binary

let me know if the webpage failed to convey this.

OK, sure. I don't see the usability of installation as separate from the usability of software, but it can be argued both ways. I think if that gets fixed we can just drop the parts about it being for tech-savvy people entirely.

Agree we can make it clear the reviews are independent. Amir?

For what it's worth, there's a reason I'm so focused on the install. At Google we measure how many users make it from "see product web page" to "running the software". It's amazing to see how people drop off through this funnel. There are some interesting findings from these studies:

• Users abandon installs a lot. Of people who see the web page, typically only a small percentage will actually end up running the software.
• Literally every additional click required kills a significant number of installs. Users really do give up at every possible point through an install. Google has put a ton of effort into reducing the number of clicks needed to install our software. If you have Google Update already active on the system, the number of clicks on Windows to go from the Chrome web page to it being open and running on your desktop is tiny - actually there are no confirmation prompts at all, IIRC.
• The longer an install takes, eg because of a large download, the lower your success rate will be.
• If anything is even slightly confusing, that will cause users to drop out too. For instance the way MacOS X handles software installation by default is consistent, GUI oriented and totally confusing (I'm talking about the DMG + drag/drop to apps folder). There's a reason virtually all consumer software has a link to the Applications folder with a giant arrow telling you what to do, and it's because huge numbers of users won't be able to correctly install the software if you don't do that, despite that it works exactly the same way for almost every MacOS app. Apple really dropped the ball on the design of this setup.

So when I see an install process that involves manual dependency resolution, what I see is not a minor picky issue but a massive drop in the success rate of users who click through that. Like probably >90% abandon rate. And if users click through to this client and are overwhelmed by the instructions, or fail to correctly follow them, they won't necessarily go back and try a new one. They'll just drop out entirely and we'll lose them for a while until one day they decide to try Bitcoin again (assuming that ever happens).

This can largely be resolved with simple installers. The advantage of being open source is you don't require a long EULA either (this also drops the success rate). You can literally have a one click install'n'go experience. You can, if you want, beat most commercial software for ease of install.

I agree with what you wrote here; it is true that we do not have an easy installation method yet.
but your text does say that; it says that the software itself, not the installation, isn't user friendly and is more suited for tech savvy people.

in addition, I believe the page should make it clear that it is a review, and that it does not necessarily reflect the views of the client developers.

The more I think about it, the more I like Luke's design. It would have been great the first time I was picking out a client. I still don't know what all's available. Maybe it could be "perverted" to include more subjective and apparently controversial wording such as "development focus" (e.g. "slow, widely-tested, conservative development cycle," "easy-to-use," "requires minimal computing/network power"), "advantages" ("allows 'offline wallets,'" "doesn't require 8+ hours of downloading and verifying a blockchain before usable," "can be used anywhere you have web access without requiring downloading anything"), disadvantages ("Requires at least 2gb of RAM," "minimal functionality doesn't allow you to do much more than send and receive bitcoins," "doesn't support encrypted wallets," "proprietary software disallows community from thoroughly examining the software").


Any jargon can link over to a FAQ section which explains the basics of downloading the blockchain (advantages & disadvantages, and how a Bitcoin wallet is able to function without one), what an offline wallet is and how it's able to function, what a "QR code" is and how some Bitcoin clients take advantage of the technology -- stuff like that. Mostly facts with some minimally-controversial opinion thrown in. I'd like to see "typical RAM usage" included in the fact sheet, too.


ETA: Then again, I like the Bit-Pay design, too. It's extremely simple, but it's so populated (even while still excluding quite a few clients), it seems pointless to have one "intermediate" option and then a giant pile of clients in the "advanced" section where it seems explanation differentiating them is necessary.