We have several projects/products mostly for B2B customers. Internally we use Confluence for documentation all sorts of decisions, processes, architecture, operations guides, research, security, high level, etc..

Our documentation can still be a mess, with many documents covering multiple aspects that makes it difficult to organize.

How do other people structure project documentation that makes it easy for not only your team to find things, but also business people, and devs from other teams so they don't bug you about "where do I find X?"

I’ve mostly been a self learner with a project based bottom up approach, which sometimes leads to knowledge being siloued from each other. So I decided trying to bring all newly acquired knowledge in structured form in Notion, which turned out to be extremely calming as I hereby somehow don’t have the urge to recall things or spot my self as fraud because I couldn’t remember them that amplify my imposter syndrom ..

Have you tried similar things? What was your experience and feeling afterward?

Basically there is no central document which contains proper business analysis. We have Epics and stories structured in JIRA. I have a digital One Note notebook which contains all my notes but it would be confusing for an oncoming BA.

What should I do for the next 2 weeks to ensure the new person has enough information?

Although I know there's no excuse for my bad documentation, I can say this project was a mess and has insane timelines making it necessary to cut corners on documentation and deliver.

I am a software developer with a decade of experience. Recently, I made a switch to teaching and took the role of facilitating and assisting students from inner city, ghettos and underprivileged sections of society.

I have a dump of short questions with their answers and explanations on high school biology. Elements in this set of Biology question-answer have no title or tags to group them. They are in random order.

My project has two stages.

In stage 1, classify these questions into different clusters according to their topic. Which algorithm would you suggest for this task?

Stage 2 is a bit intricate. The objective is to present the questions to the students in the order that facilitates their learning. In other words, every question should be followed by a question that the student is more likely to answer.

The idea is if we present questions in this way, it will help students learn the topics at their own pace and as per their thought process. The system should learn the order by asking students questions and analyzing their correct and incorrect answer patterns.

I think the Bayesian network is more suitable for stage 2. But I was advised that a neural network is more fitting here.

What are your thoughts? Which algorithms would you suggest?

My background is extensively in audit. Outside of passing the tax section of the CPA exam (REG) and some taxation courses during undergrad and grad school, my knowledge is relatively limited when it comes to niche areas like this.

I've been watching youtube videos, and picked up some books - but holy hell they are often so convoluted and MANY of them are extremely outdated.

I assume some of you may do much of the tax planning yourself, or have a general idea of where to get some of this knowledge in recent years.

Any suggestions for books/podcasts/or online videos that will be properly in depth for a CPA trying to learn this?

Hi guys,

I've been managing the build system for a front-end app at my work for the last few years, and a recurring issue between my team and others in the company whom we work closely with is documentation: either not being able to find some "document" someone threw together after a question came up one too many times and sent in an email some months ago, finding something that was good a couple years ago but the relevance of it has succumbed to the ravages of time, or a complete lack of the document in question.

My team is small and gets random questions from PMs, a dev from one of the back-ends we interface with, customer success reps, etc. throughout the year and it often results in long internal email chains which lead to, "I think X answered this a while ago, he did a little write-up on it; let me see if I can find it". These notes just end up splayed between the shared drive, inboxes, one of the products' download sites... and my Team Lead's tried to resolve it before by setting up a "central" knowledge base, but that just gets forgotten and added to the pile of past "central knowledge bases" you have to sift through to find something a year later.

Has anyone run into a more elegant solution to this bloat out in the field or have any anecdotes?

Thanks

I know there are A LOT of posts like this probably because documentation means a lot of different things in different situations. I suspect that a lot more of what I am looking for is possible but I have not used/spun up every possible tool and looked at every possible setting on those tools. However, I did some research and I suspect I will stick with Wiki.js (current leader :) )

Features:

1. What you see is what you get (WYSIWYG ).
   Context: My user base is non technical and if you can edit right on the page with tools as close to how the original page will look, also with minimal navigation to the editing interface (i.e. single "edit" button right on the page, think google sites). So, no markdown, no wiki formats, no wordpress navigating to the right page to edit. KISS above all else
2. UTF-8 first.
   Context: Documentation will be in language that requires latin-extended charset. This is sort of no longer an issue, however, it sometimes creeps up with less mature products and it's a critical requirement, hence the mention.
3. NO login edit. (OPTIONAL)
   Context: The system is in a local network and there is no difference in who edited or added what. Only real user I need is admin and any other editor does not matter. Only thing that I may would like to have control over is if we open the docs to outside people coming from public networks could not edit it, however, its a very remote consideration. Other option is to have a login prompt going out to active directory after you press an "edit button" not content viewing (but have the edit button on screen).
4. Docker deployment.
   Context: Do I need to explain why here? :D
5. Low resource use. (OPTIONAL)
   Context: Traffic to is going to be 5 users one a month at maximum, so it may as well store data on drive without a database. It would be nice to be more "green/energy efficient"

Contenders:

KING OF THE HILL: Wiki.js has most of that I need however WYSIWYG is functional but a little limited compare what can be found in notion or other modern UIs and I could not find an option for "login less" editing (guest permissions for editing is grayed out)

Bookstack: Close second might switch to it but it does not have LDAP support.

Outline: Slick look rich features disqualified due to need of external auth

Confluence: Disqualified due to ending support

Docusaurus: Disqualified due to markdown requirement

Dokuwiki or Mediawiki: Disqualified due to need for wiki editing syntax

Have I missed

20 years ago i discovered i should've been a professor. Now in my 50s, I'm trying to figure out how to teach on a variety of topics but i can't stop with desire to learn more. Started using Obsidian to help link thoughts and knowledge together, but struggling to know how to make this usable to anyone that thinks linearly.

Does anyone know a good counterclaim that could be given because my teacher said all my objects basically just repeat the same thing so I should try finding a counterclaim but can't think of anything, any ideas?

Welcome to the Theta Network subreddit! Please familiarize yourself with the subreddit rules and start here if you have any questions about Theta.

The following links should provide you with more information about the technical capabilities of the Theta Network as well as keep you up to date with latest developments of the network:

• Theta Website - Main website for Theta Network
• Theta Documentation - Central location to learn everything about the Theta Network
• Theta Whitepapers - In-depth documentation and materials about the Theta technology stack
• Theta Knowledge Base - Specific articles relating to various issues and processes relating to Theta Network
• Theta Network GitHub - Protocol code repositories, SDKs and more
• Theta Explorer - Explore the Theta Network blockchain blocks, transactions, stakes and more
• ThetaNetwork@Medium - Latest news and updates on developments from Theta Network
• Official Theta Network Discord* - For support and discussions relating to Theta Network
• GuardianMonitor (community run website) - Guardian Node statistics, staking earning estimates
• NFT Minting tutorials - Community-made tutorials for minting NFTs on the Theta Network. Available in English, Spanish and French. Credit: u/sersi29.
• KnowYourCrypto #40: Theta - An Overview - Community-made overview of Theta, part of a larger series of cryptocurrency overviews. Credit: u/madpanda94.

You can also email Theta Network at support@thetatoken.org if you can't find an answer to your question in one of the above links.

* IMPORTANT NOTE ON DISCORD: It's highly recommended to turn off Direct Messages from Server Members on Discord in order to minimize the risk of financial loss due to scammers pretending to be staff. Here is a guide on how to do this. Remember: Theta Network staff or support will never initiate contact

Known news : Veoneer sold to Qualcomm.

What's different about this ? Deal structure is different.

Source: https://www.ft.com/content/254d8ce3-de74-43f5-8cb9-3bbaa23d8c84

Read the article. It's very interesting.

How an unknown buyout firm pulled off a $4.5bn deal with a novel structure

Former Lazard banker Antonio Weiss teamed up with Qualcomm to acquire Swedish auto parts group.

https://preview.redd.it/v4ajwr9v81681.png?width=700&format=png&auto=webp&s=d3f3ddc294f458e2e1982970c7ba0659dc10e0da

What stands out about the deal is its structure. The auction for Veoneer was won in October by a start-up New York investment firm called SSW Partners. The firm had never executed a transaction, had raised no dedicated buyout fund and maintained only a bare-bones website.

The structure of their pending agreement with Veoneer has drawn the attention of the mergers and acquisitions industry. It reimagines how big corporations buy pieces of companies that would otherwise need to be acquired as a whole, deploying the bounty of private capital sloshing around the world while keeping financial disclosures to a minimum.

Qualcomm found itself in a jam this summer before the SSW arrangement took shape. The California company was interested in Veoneer’s autonomous vehicle software business with which it already had a joint venture. But the Sweden-based company had in July agreed to sell itself for $3.8bn to Magna International, a car parts supplier.

Making a rival bid for all of Veoneer made little sense to Qualcomm as it had no interest in the remaining divisions. Yet Veoneer preferred to be sold in full.

I am just thinking .... :)

https://www.otcmarkets.com/filing/html?id=14976164&guid=26CnUahj0Iglc3h#EX10-1_HTM

FOMO CORP. (“FOMO”) is providing documentation of its acquisition target SmartGuard-Solutions LLC selection by a branch of the United States Military for repeat orders of UV light disinfection fans on a non-competitive basis

As a common and (sometimes) proven belief, deep learning systems seem to learn uninterpretable representations and are far from human understanding. Recently, some studies have highlighted the fact that this may not always be applicable, and some networks may be able to learn human-readable representations. Unfortunately, this ability could merely come from the fact that these networks are exposed to human-generated data. So, to demonstrate their ability to learn like humans (and not that they are simply memorizing human-created labels), it is necessary to test them without any label.

Following this idea, the DeepMind and Google Brain teams, together with the 14th world chess champion Vladimir Kramnik, studied their creature AlphaZero from this point of view. AlphaZero is the descendant of AlphaGo, the super neural network that beat the world champion Lee Sedol in a best-of-five GO match, a turning point in the history of deep learning, as can also be seen in the wonderful Netflix documentary AlphaGo.

Unlike AlphaGo, AlphaZero is trained through self-play (i.e., it learns to play competing against itself) and masters not only GO but also chess and shogi. This trait makes AlphaZero the perfect case study to explore this idea. Moreover, given the fact that it performs at a superhuman level, understanding its functionality is also particularly useful for highlighting unknown patterns which have never been discovered by chess theorists.

Full Paper Summary by Leonardo Tanzi: https://www.marktechpost.com/2021/12/16/understanding-alphazero-neural-networks-superhuman-chess-ability/

Paper: https://arxiv.org/pdf/2111.09259.pdf

https://preview.redd.it/e8k5v5ml10681.png?width=808&format=png&auto=webp&s=6e138ed200bcb0e2f930807d490f77cc47f18b07

Well my fellow person in academia: if you feel not upto the mark, or being judged. Or not able to keep up? Read Grothendieck.

And if doesn't help, read him again. :)

Happy Christmas.

Forgive, forget and keep your deadlines aside. And in case you aren't beside your family -- this day, indulge in something, that you innocently loved about your subject.

We are currently using confluence for our internal documentation. It's been getting out of hand a bit with random spaces amd pages being created in random places. We have been managing so far because the search is ok however I want to spend some time to structure it easily for everyone to use. The one thing we do seem to do well those is use a single page for how-to articles.

I am after some ideas or examples of how you structure your documentation so that it makes sense and is easy to navigate.

I just want to mention that this is only accessible by IT folks and isn't accessible to staff.

So I have been thinking about the recent acquisitions and how this can possibly help Xbox Game Studios evolve. By the time the Zenimax acquisition closes, Microsoft will have more than 20 studios with some of those even having multiple teams. I would like to share some thoughts and maybe this can lead into a discussion.

Having worked for two international IT companies with even more employees than Microsoft, I know that knowledge sharing is key in big companies. Even without the new acquisitions, I would say XGS have one of the technically most skilled studios in the business with The Coalition. While some of you may chuckle because Gears is no revolutionary game, the effort they put into Gears 5 is amazing and it is basically a show case for a modern Unreal Engine game. They are also very good at sound design which is something that has been neglected by the industry in the last couple of years imo. Other than that I think Ninja Theory is one of the best studios when it comes to Motion Capturing. I have yet to see a game that is substantially better in terms of Mocap than Hellblade. Hellblade II seems to be even better. I also think that their Mocap expertise was a huge part of why Microsoft acquired the team. Microsoft offer something apparently called career jams where developers can connect to other developers. This doesnt have to be a developer with similar tasks and responsibilities but could also be something totally different. So what I am theoretically seeing here is huge potential for all the people within XGS to grow their expertise. Animators can learn from other animators, sound designers from other sound designers and so on.

This leads me to another potential advantage: flexible work. So Microsoft has recently allowed permanent home office work for their employees because of the worldwide situation. This on the other hand would make it possible for employees to switch to other studios. Maybe not permanently because this would mean another contract, but project-wise. This would also counter the argument that people working on one franchise all the time (eg 343i employees working only on Halo) get bored after a while. With Microsoft allowing home office work, having basically all the productivity tools and being very international and inclusive, I see no issue here. So someone bore

Just need one more object please help...

As a common and (sometimes) proven belief, deep learning systems seem to learn uninterpretable representations and are far from human understanding. Recently, some studies have highlighted the fact that this may not always be applicable, and some networks may be able to learn human-readable representations. Unfortunately, this ability could merely come from the fact that these networks are exposed to human-generated data. So, to demonstrate their ability to learn like humans (and not that they are simply memorizing human-created labels), it is necessary to test them without any label.

Following this idea, the DeepMind and Google Brain teams, together with the 14th world chess champion Vladimir Kramnik, studied their creature AlphaZero from this point of view. AlphaZero is the descendant of AlphaGo, the super neural network that beat the world champion Lee Sedol in a best-of-five GO match, a turning point in the history of deep learning, as can also be seen in the wonderful Netflix documentary AlphaGo.

Unlike AlphaGo, AlphaZero is trained through self-play (i.e., it learns to play competing against itself) and masters not only GO but also chess and shogi. This trait makes AlphaZero the perfect case study to explore this idea. Moreover, given the fact that it performs at a superhuman level, understanding its functionality is also particularly useful for highlighting unknown patterns which have never been discovered by chess theorists.

Full Paper Summary by Leonardo Tanzi: https://www.marktechpost.com/2021/12/16/understanding-alphazero-neural-networks-superhuman-chess-ability/

Paper: https://arxiv.org/pdf/2111.09259.pdf

https://preview.redd.it/t4mjebrm10681.png?width=808&format=png&auto=webp&s=58fcc96e8b1ae92469c26820528d0a5b31514365

As a common and (sometimes) proven belief, deep learning systems seem to learn uninterpretable representations and are far from human understanding. Recently, some studies have highlighted the fact that this may not always be applicable, and some networks may be able to learn human-readable representations. Unfortunately, this ability could merely come from the fact that these networks are exposed to human-generated data. So, to demonstrate their ability to learn like humans (and not that they are simply memorizing human-created labels), it is necessary to test them without any label.

Following this idea, the DeepMind and Google Brain teams, together with the 14th world chess champion Vladimir Kramnik, studied their creature AlphaZero from this point of view. AlphaZero is the descendant of AlphaGo, the super neural network that beat the world champion Lee Sedol in a best-of-five GO match, a turning point in the history of deep learning, as can also be seen in the wonderful Netflix documentary AlphaGo.

Unlike AlphaGo, AlphaZero is trained through self-play (i.e., it learns to play competing against itself) and masters not only GO but also chess and shogi. This trait makes AlphaZero the perfect case study to explore this idea. Moreover, given the fact that it performs at a superhuman level, understanding its functionality is also particularly useful for highlighting unknown patterns which have never been discovered by chess theorists.

Full Paper Summary by Leonardo Tanzi: https://www.marktechpost.com/2021/12/16/understanding-alphazero-neural-networks-superhuman-chess-ability/

Paper: https://arxiv.org/pdf/2111.09259.pdf

https://preview.redd.it/096omb8m10681.png?width=808&format=png&auto=webp&s=c375a2bfffc4949399e17c7ebbe2e2c334a2a44d