Hi riddim, great work on the Friends dweb app btw.
I’m a total beginner on here (lurked a couple times over the years) and am trying to test out the apps, I can connect to services using either the Terminal or happybeing’s handy GUI client, but I can’t seem to get the services themselves to properly run.
For instance, I have accrued enough ANT to create a profile but can’t seem to get past the “Choose searchable public identifier (optional)” step, it is apparently initializing the new profile but never progressing.
Similarly Josh’s Atlas never successfully loads (I can get the htmla nd css to load, but the service itself “fails to initialize Atlas”.
I think fees have been a bit increased not long ago; could have been payment/upload errors due to too high fees - maybe you were unlucky and need to just start over again
Have you attempted to build in a quoting facility for what fees are “currently” required to create the profile? It seems that would be quite useful for such a dynamic point of entry.
I noticed that MM sees my wallet as holding 1.019.. ANT but in the first page on Friends it displays what appears to be my ANT as 0.1019..
At first, I thought this was the price, but now I realise it is probably my balance from the perspective of the service.
In addition, I don’t have any ETH, though I have managed to get some SepoliaETH for whatever that is worth?
Working on solutions such as Friends requires not only coding but also a lot of thought and consideration of the team’s strategy for developing Autonomi. If @riddim does a lot of programming, conceptual, analytical, etc. work that is not supported by MaidSafe’s current development plans, his work will be wasted, as has happened so many times with other programmers.
Friends is developing very well, but it is not just a regular Android application; it requires many factors to be taken into account because it is being created from scratch on the “uncharted territory” of Autonomi.
oO what do you mean by this - dweb doesn’t support passwords for wallets - right?
hmmm - seconds up to a minute or so I guess; but if fees spike high and/or some other error occurs this might just not work on first attempt … simply retrying should resolve the issue
I don’t remember even typing the private key, could dweb have used the autonomi api or something, it just found the wallet and asked me to type the password when creating the scratchpad.
Today it worked straight from the first try, everything is smooth, don’t know why.
Friends seems to work great when I tried it with Ubuntu on windows 11 through wsl. I’ve been trying different autonomi apps the last days and it has not been a great user experience for several of them from the perspective of a regular ordinary person. It would be nice to be able to make promotion materials and similar to help people find Autonomi and the apps that are being created but it is not possible with the current user experience of many apps, writing commands in cli, building from source with cargo, several steps to open an app and so on.
I have tried to make a launcher today for Friends as a demo of a vision, as friends seems to work well. The launcher should work with Linux and Windows. I’ve tried it in Ubuntu through ubuntu wsl in win11 as a python app and also made an .exe for windows, even if I at the moment can’t get any app or ant-cli to upload or download in win11 as there are some connection problems. But I want to at least give an early demo taste and maybe inspire developers on the forum to make install/launchers bundled with their apps.
The goal is to inspire devs to make autonomi apps as user friendly as possible so it becomes possible to start promoting more apps to the general public and also reduce barrier to entry for trying and testing apps.
As the launcher is only a demo then it has not yet been packaged to be a double click friendly experience on linux distros. It will need, requests, tkinter, psutils modules and maybe others to run but hopefully it can give a taste of what it is suppose to showcase. On windows it should work by just running the .exe
ANT download adresses:
friends-launcher-v7.py: 17312802048149503522
friends-launcher-v7.exe: 10696411043148271040
modules used:
Summary
import tkinter as tk
from tkinter import messagebox
import subprocess
import webbrowser
import threading
import time
import platform
import requests
import socket
import os
import psutil
I uploaded these as private files, don’t know if the links will work or how to share/find the public map for them in a correct way, tell me if the download adresses don’t work.
Here is a video that showcases the friends demo .exe launcher on windows.
I strongly believe installers/launchers like this will make apps a better experience for the general users and help us reach a larger general public, because the apps are great but they also only need to be super easy to install and use, no cli, building from source, using with cli commands and so on. Easy to isntall/use I very strongly believe is 50% of an apps success. Give a heart if you agree or tell me what you think. I plan to move this launcher and discussion to a separate topic next, soon.
Hi tobbetj, yeah that sounds about right for the first message.
Once the dapp has fully loaded, messages are a bit more seamless, I’ve had a whole conversation with riddim that felt as smooth as you would expect.
I can already see a bottleneck on startup as it recovers message history (especially the more the dapp is used). Not sure what the long-term solution would be outside of cookies and network efficiency boosts, I guess we’ll see how things improve over the coming weeks/months.
yeah … and the solution with scratchpads is not the most idiot-proof store for e.g. list of friends … i think I need to change that to a dweb history type with encrypted content to get a nicer user experience … I somehow emptied my friend list a few times by now by accident …
that whirl of queries on startup could be spread out across time better and the sometimes longish delay to get message history of specific chats back could be improved by using one buffer-pad that spreads across multiple chats (and buffers the last 100 messages of us and our friends to enhance user experience) … so that doesn’t worry me too much … but for sure a lot of work left to make it a proper product …
next steps will be dweb history datatype and graphEntry exposure via API I think … without those the points up my priority list can’t be tackled …