BitShares Forum

Main => Technical Support => Topic started by: lovejoy on January 27, 2015, 03:28:05 am

Title: robohash discrepancies
Post by: lovejoy on January 27, 2015, 03:28:05 am
I'm having disagreement between my wallet, bitsharesblocks, and the robohash.org website...
Some robohashes in my wallet are different than they used to be, and some are the same.
Many disagree with the robohash.org website altogether. I'm running this wallet on v0.5.3

What gives?

Here is one example:
(http://bitscape.io/wp-content/uploads/2015/01/robohash_whaaa.jpg)

In this case, the bottom one from bitsharesblocks is most familiar and my wallet displays one i've never before seen.

I no like. 
Title: Re: robohash discrepancies
Post by: lzr1900 on January 27, 2015, 03:29:43 am
I have the same problem as yours.
Title: Re: robohash discrepancies
Post by: xeroc on January 27, 2015, 07:17:17 am
<- that thing changed over night :( should be violet and fat 8)

It's a link to robohash.org

edit:
(http://i.imgur.com/hDY4YOi.png)
Title: Re: robohash discrepancies
Post by: mira on January 27, 2015, 07:26:56 am
<- that thing changed over night :( should be violet and fat 8)

It's a link to robohash.org

^^I still see a violet and fat thing here by your forum name. ..?  When I type 'xeroc' in at robohash.org I see a white thing with yellow eyes.


In the client, my robohash has changed only on the account overview page, is the same old one on other pages...


Edit: Now it's changed. What is up with that!!
Title: Re: robohash discrepancies
Post by: arhag on January 27, 2015, 07:28:12 am
Aghhh, why are we still using robohash / SafeBots?

You should either make transactions with people explicitly added into your contacts / address book OR transactions with important information like the recipient (and perhaps memo) automatically filled out by clicking on a link on your computer.

The only time you should need to manually be absolutely sure you are dealing with the correct account name is when adding the contact to your address book and giving it an alias. And that is only when you can't have the contact addition be automatically done for you with mobile-to-mobile communication, meaning in situations when someone verbally told you their account name. And in situations where you verbally communicate your account name, are you really going to describe your SafeBot? A 3 digit checksum is far more practical for that purpose.
Title: Re: robohash discrepancies
Post by: xeroc on January 27, 2015, 07:37:41 am
A 3 digit checksum is far more practical for that purpose.
let's use a RS-code instead that allows to fix up to 2 typos per account name

Checkout NXT, they are doing it that way
Title: Re: robohash discrepancies
Post by: arhag on January 27, 2015, 07:42:59 am
A 3 digit checksum is far more practical for that purpose.
let's use a RS-code instead that allows to fix up to 2 typos per account name

Checkout NXT, they are doing it that way

Sure, whatever, sounds good. The point is that communicating your natural language compatible account name should be something rare. It should be reserved for exchanging contact info between people in person who don't have their smartphones handy (or don't want to bother to use it), or for people broadcasting their account name (similar to how people might announce their Twitter handle) over mediums in which some acceptable level of biometric authentication (voice and/or face) is possible, such as TV, radio, audio/video podcasts, etc. It shouldn't be used for any other situation.
Title: Re: robohash discrepancies
Post by: graffenwalder on January 27, 2015, 11:12:59 am
Should a ticket be made at bter, to change their avatar on the deposit page?

Or should we wait until we find a solution to the changing avatar problem?

Edit:

Just made a ticket to give them a heads up
Title: Re: robohash discrepancies
Post by: liondani on January 27, 2015, 01:19:00 pm
same problems with my wallet,  I have new robots faces now,  what a mess  >:(

Sent from my ALCATEL ONE TOUCH 997D

Title: Re: robohash discrepancies
Post by: svk on January 27, 2015, 01:25:26 pm
This could just be an issue with the wallet sending an incorrect string to robohash.org. My robot still matches but I haven't checked in the wallet. How about the robots on bitsharesblocks?

Just noticed the OP includes bitsharesblocks. Strange, I'll look into this some more once I get home. The wallet btw uses: http://robohash.org/set_1/robot_name while I believe I'm just using http://robohash.org/robot_name.

If you check http://robohash.org/set_1/media.bitscape it's the same one that's in the wallet.
Title: Re: robohash discrepancies
Post by: xeroc on January 27, 2015, 01:30:36 pm
This could just be an issue with the wallet sending an incorrect string to robohash.org. My robot still matches but I haven't checked in the wallet. How about the robots on bitsharesblocks?
Nop .. My avatar is a link to robohash that hasn't changed for weeks .. though the image changed .. it's definitely on the robohash.org side
Title: Re: robohash discrepancies
Post by: svk on January 27, 2015, 01:34:57 pm
Think I found the issue on the bitsharesblocks end, I'm fetching http://robohash.org/dev.bitsharesblocks.png, I need to remove the .png and add /set_1/
Title: Re: robohash discrepancies
Post by: svk on January 27, 2015, 01:36:00 pm
This could just be an issue with the wallet sending an incorrect string to robohash.org. My robot still matches but I haven't checked in the wallet. How about the robots on bitsharesblocks?
Nop .. My avatar is a link to robohash that hasn't changed for weeks .. though the image changed .. it's definitely on the robohash.org side

Looks like they changed the hashes for set_1 then, I'm pretty sure they used to be the same as the standard one because the robots on bitsharesblocks were the same as the ones in the wallet.
Title: Re: robohash discrepancies
Post by: graffenwalder on January 27, 2015, 02:05:15 pm
Something strange going on now.
My profile pic is linked to robohash. But if you keep refreshing the page, it changes from the old to the new avatar, back and forth.

It only happens with the 150x150
http://robohash.org/set_1/graffenwalder?size=150x150
The 40x40 seems to stay the same
http://robohash.org/set_1/graffenwalder?size=40x40
Title: Re: robohash discrepancies
Post by: valzav on January 27, 2015, 02:08:25 pm
yes, it looks like robohash.org has changed something on its side
Title: Re: robohash discrepancies
Post by: abit on January 27, 2015, 03:54:41 pm
This could just be an issue with the wallet sending an incorrect string to robohash.org. My robot still matches but I haven't checked in the wallet. How about the robots on bitsharesblocks?
Nop .. My avatar is a link to robohash that hasn't changed for weeks .. though the image changed .. it's definitely on the robohash.org side
Your new avatar looks strange.  ???
Title: Re: robohash discrepancies
Post by: abit on January 27, 2015, 03:59:42 pm
This could just be an issue with the wallet sending an incorrect string to robohash.org. My robot still matches but I haven't checked in the wallet. How about the robots on bitsharesblocks?
Nop .. My avatar is a link to robohash that hasn't changed for weeks .. though the image changed .. it's definitely on the robohash.org side

Looks like they changed the hashes for set_1 then, I'm pretty sure they used to be the same as the standard one because the robots on bitsharesblocks were the same as the ones in the wallet.
This looks same to my original robot: http://robohash.org/set_4/a.delegate.abit
Title: Re: robohash discrepancies
Post by: svk on January 27, 2015, 06:02:09 pm
I've pushed an update to Bitsharesblocks that should keep the robots consistent with the wallet. As graffenwalder noted, there's definitely an issue with the ?size param, there appears to be at least two different robots generated depending on the size you use, so for anyone looking to use this, be careful with the size!
Title: Re: robohash discrepancies
Post by: lovejoy on January 27, 2015, 06:12:23 pm
Thanks to the crack BTS detective team for getting to the bottom of this and pushing updates / clearing confusion. Especially happy it wasn't just me! :)

I agree there are perhaps better solutions, but for now we're on the Robohash / SafeBot tip, so good to keep it working as intended.