A loosely moderated place to ask open ended questions
If your post is
- Open ended
- Not offensive
- Not regarding lemmy support (c/lemmy_support)
- not ad nauseam inducing (please make sure its a question that would be new to most members)
it’s welcome here!
- 0 users online
- 1 user / day
- 31 users / week
- 135 users / month
- 410 users / 6 months
- 8 subscribers
- 1.2K Posts
- 12.5K Comments
This is actually a funny question if you think about it a bit more.
There is no such thing as a centralized system globally speaking, at least not in the real world. Having global centralized systems is actually a very new thing that only came to be on the internet.
Otherwise the world is by necessity deeply decentralized. Even at the level of individual nation states it usually is much more decentralized that it appears at first. Humans simply can’t do centralized systems, this only came to be once we had a global computerized network.
The postal system.
I get that there are private shipping companies and countries, but having a national publicly owned postal system that is coordinated in a centralized way is much much more efficient than any decentralized system could operate for distribution of physical packages.
This is a good example for a natural monopoly, but that does not mean that there can’t be regional suppliers that cooperate with each other in a decentralized way or covering different market segments.
Could, but should? Depends on the goal we’re trying to achieve. Efficiency of delivery I’d say monopoly in this context. Regional specialization maybe decentralized, but what benefits in postal services would it bring?
What benefit would centralization bring over a regional monopoly service? Postal services have regional distribution centers anyways, so you just agree on a inter-operation standard and then these can be independent entities. I see absolutely no increased efficiency or other advantage of some kind of centralization in such a context once every entity agrees on such a standard.
Edit: It would simply work via the typical concession model. Every 5 years of so the government or who ever decides the interoperation standard makes a open tender for a certain regional service or a specialization topic. Companies or cooperatives can submit their offers and proposed cost recovery models and the best offer gets a natural monopoly for 5 years (or 10 years what ever). Such a model is commonly used and works quite well if it doesn’t involve big investments in long lived infrastructure.
Maybe we have a different understanding of centralization? For me centralization means a centrally controlled and monopolized service. Just having a natural monopoly in certain areas or topics doesn’t make a service centralized, it also needs to have a single entity controlling everything.
Isn’t the US postal system a single entity controlling everything? Yes they have regional distribution centres but it’s still a single entity. I guess it depends on your definition of everything. I don’t mean a single distribution centre of course.
Agree that we’re using different definitions of decentralization. My interpretation of decentralization is where transfer of authority or management of a service or function moves from a centralized national or digital single-entity/corporation to a geographically regional or digitally autonomous (multiple) entity(ies).
IMO natural monopolies are good examples where decentralization of processes should not occur. However, that’s not to comment on whether those centralized monopolies will be more resilient to political opportunism or other potential negative impacts of centralization. This is a layered concept for sure, easy to get into debates about.
Healthcare for example, government should provide gratis (free as in price) healthcare for everyone
This is for the funding, but the healthcare system (and tools) can (and in my opinion should) be decentralized.
Even the funding could be partially decentralized (and when there will be a worldwide wealth redistribution system, it should be decentralized, in my opinion) We can imagine a system with an universal income for every human ( for it’s everyday life, and funding for medical devices and drugs asked by hospitals and practitioners to their local healthcare organization, which is forwarded to a committee of the united world healthcare funding system if there is a lack somewhere.
Agree with regard to centralized funding but decentralized and autonomous practice.
My experience comes from working as a medical assistant and nursing student in the US.
Medical records. When a patient seeks care from a new provider the practitioner often needs access to their complete history to treat conditions that may appear trivial at first glance, but might be a symptom of something more serious.
The current medical record system in the US allows for a patient to hop from care provider to care provider, selectively disclosing their medical history so that no single provider has a complete medical record. This design has been preserved from a time before a universal medical records were possible in order to address valid patient privacy concerns, but from a providers perspective it makes practicing medicine far more dangerous.
If a patient presents for a sinus infection and is prescribed a penicillin derivative without ever disclosing that they are allergic, and then the patient dies of an anaphylactic reaction, the patient’s family is likely to sue for malpractice. Did the patient simply forget to mention they were allergic or did they decide their last reaction was so long ago that they were no longer allergic? It doesn’t matter, the family’s lawyers will scrounge for a medical record, or several, documenting the penicillin allergy. Even if the practitioner’s documentation of the visit is perfect, and the patient’s family loses the case, it’s still a soul-crushing and expensive experience for everyone involved.
In another situation a patient with an opioid addiction may visit many providers in their region in order to receive multiple prescriptions for painkillers, this is very common.
Without a seamless medical record system these and other preventable outcomes will continue to cause misery to patients and providers. I believe a central medical record database would drastically improve the quality of healthcare in the US.
Decentralization of medical records is a single factor contributing to the fact that US doctors and nurses commit suicide at a rate about three times higher than the general population.
While I’m on the soap box: Medicare for all is the single most important change to push for. If you live in a country with socialist healthcare, please protect it.
Couldn’t medical records be normalized (standardization of data structures), but the data hold by citizens or local administrative governments (and later in a decentralized storage system¹ ? (with e2e encryption and access control by the holder or some emergency mechanism)
This way, you wouldn’t have a single organization storing all the medical data (and thus having a lot of power), but every medical practitioner or organization could access medical records the same way and common software could be developed
¹disclaimer: not blockchain, I think blockchain is irrelevant for most applications
I think a centralized database is the simplest and most politically feasible with current technology. In theory users could hold all the data on their own device, but this introduces additional complexity and problems. For instance, many of the notes that are included in a medical record are not meant to be seen by the patient e.g. “At presentation patient stated pain in lower back, then stated the pain was in their upper back, then stated pain was in left leg. Inconsistent report, pain medication not prescribed. Patient’s Medical Record Number flagged for future providers to be aware of possible substance abuse”
If the user holds this data themselves, then they will presumably be able to read that data, and that will make providers jobs that much more uncomfortable knowing the patient will likely read the sensitive information that needs to be stated, but has a high likelihood of making patients uncomfortable and angry. If the user holds this data they will also be able to edit and delete crucial parts of their medical record and without a proper medical record providers can not administer proper care.
What is currently happening is that Electronic Health Record software companies are starting to build these centralized databases / make sharing easy between health systems using the same EHR software e.g. EPIC ShareEverywhere.
I believe a publicly accountable institution should be responsible for managing a central repository, not a private company, and I believe that at the current time a centralized collection of medical data is possible while a decentralized, but seamless, database has yet to be proven.
In my country (france), such a service is currently deployed by the public healthcare organization. After the digitization of public administrative services by every state (which is mandatory in the european union), the EU will surely make it mandatory to have interoperability between the different health systems.
What is nice with this approach is that each state keeps control over their citizen’s data, but the mandatory interoperability (for administrative documents, and hopefully healthcare later) will make it easy for every citizen to move from one country to another.
It will be federated rather than decentralized, though, and I think that it should become way more decentralized so that every citizen really have a control over their data. I don’t think we can trust governments of nations. I agree that a reliable decentralized database approach doesn’t exist yet.
For the issue of “records that is not meant to be seen” I don’t think that it should exist. I think everybody have the right to be informed about their health¹. You can forbid modifications of the medical record by using merkle trees and signature, but I don’t see any reason to hide this information from the patient. For the example, if the person is told that they have the same comportment as an addict, don’t you think they will be more inclined to seek for help if they really have an addiction issue ? (provided this help is free and accessible, of course)
¹ if the person is recognized as responsible of themself
Applications that rely on a central authority. For example, the current NFT artwork mania runs on a couple of centrally-operated websites. Once those sites go down, most of the URLs that those NFTs are linked to are going to stop resolving. It’d be way cheaper and environmentally safer to just use a centrally-operated RDBMS (disclaimer: I think NFTs and most cryptocurrency stuff is absolute bunk).
I still don’t get why authenticity is important in art. In my opinion, copying art is something great to share knowledge and culture.
I think NFTs could be a great way to fund arts or charity. Especially if they can smart contract money to their cause with each transfer.
I personally don’t know how NFTs could really work as an investment mechanism because they aren’t a currency and their ties to artwork are about the same as reddit posts with the same URLs. If there isn’t an agreement under law giving people exclusive rigbts or monies from their NFTs then they are even more speculative than physical art.
Not saying they couldn’t work, just that I don’t comprehend the mechanism keeping people buying them at higher and higher prices.
Holy shit, you’re right: most NFTs are just URLs! I thought the images themselves, or at least the hashes of the images, would be stored in the blockchain. The stupidity of altcoin developers never ceases to amaze.
I wouldn’t say they are stupid, they achieved what they wanted, namely making a shitton of money
Irrelevant, but just thought I’d say: If it’s impractical to decentralise, it should be owned publicly e.g. open source, cooperative, mutual or government business model
Electrical grids (not generation)
Huh? Electrical grids are decentralized in many ways right now.
Yes but the question is what should not be.
…and the are decentralized for very good reasons…
Maybe you could list our a couple of the good reasons? In Canada the provinces that have centralized electrical grids have the cheapest electricity while provinces with decentralized have the higher power costs. Passive aggressive comments are sometimes entertaining but usually they come off flat when there’s no empirical evidence to back them up. 😉
Electrical grids across the world have been decentralized since the beginning - hence the name - for various reasons. Resilience: a long distance line can be disconnected due to a fault and the network need to be able to survive the fault without leaving users without power. Maintenance: you need to be able to disconnect the line without cutting power to users. Economics & corporate politics: even if generation or distribution is done by the government in many countries, the companies actually doing the work are often private. You don’t want a single company to have huge bargain power. Also: resilience from air strikes and carpet bombing (yes it’s a concern).
It’s pretty common for datacenters to be connected to multiple power lines possibly from different providers for reliability.
If this sounds similar to how the physical Internet is built it’s not unexpected.
Ahh, you’re talking about technical operations. I’m referring to governance. Yes, everything can be looked at in a decentralized way depending on the frame of reference. Good job!
As mentioned in other answers, a lot of governance is decentralized (e.g. across different countries) as well.
I think they’re saying electrical transmission should be centralized, but generation should not be. Problems with double negatives…
Yes you’re correct. As per the question asked with the “NOT”, imo Electrical Grids should not be decentralized while generation should.
But transmission isn’t centralized, and can’t be centralized. Because its purpose is to transmit power from power generators (both centralized and decentralized) to (located elsewhere) loads.
I assumed the OP meant generation because the other way around doesn’t make sense.
I may also be thinking about the physical world with the OP thinking about ownership, but that makes even less sense. Transmission and generation should be paid for by the end users. Preferably in whatever way reduces their costs without making them a burden on other participants.
Maybe the statement is talking about centralized ownership for the purposes of economies of scale. Even that can be broken down to cooperative ownership as much of the less urban areas of the United States already are.
I meant transmission, systems that maintain load balancing, etc. Generation of renewables should be decentralized, but requiring generators to comply with creating transmission lines to rural areas or cities will create a bunch of issues around who is responsible for what, standards mismatch, and extremely high cost for smaller populations. It doesn’t seem like a good idea to me.
So distributed generation shouldn’t happen? As far as I know distributed generation is happening and it is likely to increase as time goes on. https://en.m.wikipedia.org/wiki/Distributed_generation In general non-renewables should be centralized due to economies and efficiencies of scale but renewables almost have to be distributed due to the amount of space they take and localized weather effects.
Yes distributed generation is good. I said not generation in my first line. The wikipedia article you referenced doesn’t mention transmission grid infrastructure, I’m curious, who is responsible for that in a distributed grid anyway?
Transmission owners own transmission. Transmission operators operate transmission. Reliability coorfinators have a wide overview and exist to help keep the grid together.
IMO, the reliability coordinators, transmission owners, and transmission operators should be centralized while generation should be decentralized.
This reminds me, the postal system should be centralized. A decentralized postal system would be an absolute mess.
What makes a grid centralized besides generation?
The maintenance of the transmission lines, smart load balancing, maintenance, etc.
Search engines. It just doesn’t work.
SearX is decentralized meta-search engine
YaCy is decentralized true search engine (i think)
Yeah that’s an accurate way of putting it.
Has it been tried? it might require re-thinking how search engines work. If the databases were all formatted consistently, it seems possible.
Yes, there is YaCy. And it sucks. It brings like 10 results in single word searches, usually none of which are relevant. When you write a sentence of sorts it just freezes.
There’s Yacy. It worked ok when I was running it locally, but the moment I connected it to the rest of the network my search results were dominated by porn and spam sites.
My google and duckduckgo search results are also dominated by porn sites.
well hey, I won’t judge
Search engines are incredibly conplex. the problem might not be decentralization, but just proper investment and development.
the problem as I understand it is that with yacy you request relevant results from peers in addition to your own local DB. but that makes it really easy for spammers to keyword-stuff results and push their sites to the top.
I’m not sure how we can trust random people to give you search results without being overrun by spam.
Interesting, couldn’t you exclude results from servers that did this?
I agree. I think it’s sad that YaCy never worked as well as the alternatives, but it is a very cool project nonetheless!
Maybe in the future someone will try the concept again and succeed.
Also: Google is getting worse and worse, so it’ll get easier over time ;)
Backend is implied here (quasi-decentralised frontends can be great), and agreed it’s incredibly valuable to have a central index of content, rankings, answers. Maybe it can be overcome by advances in data transfer speeds and some talent taking interest in the project (for example, would centralised search be used on Mars?)
I was going to say Databases but I suppose you could have multiple sources to feed and make a local database…
orbit db says hi :)
there’s also actorDB which I think is more mature
actor db is literally abandoned though
This looks pretty cool, but is it completely P2P like OrbitDB?
Also, it looks a little bit abandoned: last commit was in 2019
Yeah, doesn’t look like it’s actively developed unfortunately. It is fully distributed though, from the docs:
What I mean is that, OrbitDB seems to be open P2P (a network that anybody can connect to), while ActorDB seems to be private P2P
oh yeah ActorDB is a private P2P
Is that production ready?
i’m not sure what concretely do you mean by that…
it’s still new, as ipfs iself is, but it’s already used by multiple applications to different degrees of success
That is, is it in a stage where it’s usable to some degree outside of testing
i haven’t used orbitdb yet, so i’m not sure what state it is in exactly
but this depends on what your application for it is and what you’re trying to replace with it, just like ipfs itself can be pretty good for certain purposes, but still be lacking in certain things like dynamic websites or streaming
i would recommend asking on their gitter
oh THIS is awesome!