Homolog.us – publish

An Interesting Observation about the Architecture of Decentralized Publishing System (iii)

Fri, 04 Sep 2026 00:00:00 -0700

This is a continuation of my earlier posts here, here and here.

I spent some time on thinking about the code architecture of the new publishing system. I have an interesting observation to share.

As I mentioned, I want to start with building something similar to an online preprint or publication server. Based on my understanding, the code can be split into two (or three) distinct activities.

1. Displaying the Paper

Displaying a paper on the webpage involves several tasks. The texts and images need to be placed properly along with correct rendering of equations. Citations need to be linked to external sources. The pages should scale correctly in phone, tablet and laptop.

2. Human Interactions

When an author submits a paper, the system sends notification to an editor. The editor likely opens a page with papers under his consideration and then assigns various reviewers to them. The reviewers send back feedback, and based on them, the editor accepts or rejects the papers. Of course, there are additional back and forth exchanges, but they are similar to the presented tasks.

In terms of system architecture, various users are assigned roles - (i) author, (ii) editor, (iii) reviewer. Most users likely have multiple roles, because they are authors for their own papers, and reviewers for others’.

3. Payment

Those are the only two activities for preprint servers. For paid journals, a third code block takes care of payments.

The main point here is that these three components can be developed independently. For example, we can model the human interaction block by using a short paragraph of text instead of an entire paper. In this model system, an author will submit a paragraph, reviewer will review it and the editor will allow or reject publication of the paragraph. In fact, the comment moderation sections of blogs and forums use such an approach.

Interesting Observation - Here is What is Missing

When you think through the current architecture, you see one piece entirely missing from it, and that is the interactions berween the users (readers, authors, even reviewers).

For example, in a publication system built for 21-st century, the authors and other readers should be able to join an online discussion room like Zoom or twitter space to chat about the paper. That discussion can be recorded so that others missing the discussion can listen to. I can think of many other interactive activities the system should naturally support, but does not. Instead what we have is a grudgingly made improvement over the book publishing system from the mid-1800s.

Decentralized Publishing System - Code (ii)

Thu, 03 Sep 2026 00:00:00 -0700

This is a continuation of my earlier posts here and here. This update will be short.

As I understand, we can dream about many different systems, but such abstract discussions are meaningless without actual implementations. Code and software architecture will ultimately decide what is feasible. Moreover that code will have to be maintained and improved over the long-term.

Therefore, as a next step, I will build a protype similar to the existing preprint or publication servers. That is the minimal baseline to start with, and then we can think of how various aspects of the process can be changed.

I will post an update, when the prototype is ready.

Architecture of a Decentralized Publishing System - (i)

Wed, 02 Sep 2026 00:00:00 -0700

The Internet revolution never truly arrived in scientific publishing. Instead, an old system that was functional in its time was scaled up in a largely cargo-cult manner, producing today’s grotesque, dysfunctional, and extraordinarily expensive publishing ecosystem.

Most scientists tolerate this absurdity because they are insulated from its costs. First, they learn to accept the dysfunction as the inevitable “cost of doing business.” Second, universities pay the subscription bills, so individual researchers working there rarely feel the price and paywalled gatekeeping directly.

I already wrote about about how the old publishing system was simply scaled up during the Internet era. The scientific societies were gradually replaced by journals and editorial boards.

Scientific societies had a strong incentive to protect their honor and reputation. Their journals were part of a scientific community they themselves belonged to. Commercial journals are different. An editor may be a dedicated scientist, but the labor of him and unpaid reviewers ultimately serves a business interest. We replaced institutions built around scientific reputation with businesses built around extracting value from scientific labor, and then pretended that nothing fundamental changed.

Flaws of the Current System

Before looking into how a better system can be built from ground-up, here are some of my specific complaints.

It is nearly impossible to track the publications of individual authors. Google Scholar is probably the best option, but it does not provide the papers themselves. Instead, you are left chasing papers scattered across the Internet with some freely available as PDFs, some hidden behind paywalls, and others simply inaccessible.

I remember that, in the early days of the Internet, many scientists created personal websites and posted PDFs of their papers. Journals promptly harassed them for “copyright violations” as if making publicly funded research actually accessible were an offense.

That brings me to copyright law. Philosophically, copyright exists to protect authors: when someone spends years writing a novel or textbook, it makes sense to give that person ownership of the resulting work.

Scientific publishing is fundamentally different. Scientists generally receive no payment for their papers. Journals nevertheless use copyright law to control and monetize access to research that is often publicly funded. This is an abuse of a law designed to protect creators. Neither the scientists who produced the research nor the taxpayers who funded it receive a penny from the resulting paywall.

My biggest complaint about the system is that it failed spectacularly during both the ENCODE saga and Covid. Politically favored papers moved fast through the system, whereas those challenging the status quo were delayed. That is anti-science in the name of science.

Fresh Thinking

I always start with this question - if the current system for sharing scientific results did not exist, what would the scientists like to build? I think they would take the best ideas from various online tools and sites in existence - github, substack, wordpress, reddit, stackoverflow, etc.

Some thoughts -

The system needs to have a clean break from current model, just like the internet was built with a clean break from the phone companies. Maybe the specifications can be discussed through a method similar to RFCs.
The system needs to be long-term viable.
The system needs to be politics-proof.
The system needs to be liberal about accepting contributions with protections from spams, hoaxes and other harmful content.

Decentralized Honor Societies

I am thinking about a two-layer system with the base layer accepting a large volume of papers, and a second-layer providing the role similar to the honor societies of the older days. For example, a small groups of researcher working on a topic can create a topical “journal” as a layer above the base system. This is the new honor group, and its reputation will be based on the quality of papers it selects.

Such “journals” can choose to review papers before accepting them. Regarding review method, I am wondering whether a public (online) talk would be one possibility. By this time, the paper is already posted in the base system, and a “journal” layer is deciding about accepting it. A presentation will serve both the purposes of allowing the author promote his work and the members of honor society vet it.

The same paper can be accepted by multiple honor societies (interest groups). This will be equivalent to cross-posting of blog posts.

Code and Finances

Whether the new system will make or break will be decided by two factors - code and finances. I will talk about them in a later post.

What Are the Features of a New Publishing Model?

Tue, 01 Sep 2026 00:00:00 -0700

Clearly the current publishing model is utterly dysfunctional. The best way to see that is to ask the following question. If the current system is thrown away, and the scientists gather to create a brand new system for publishing, will they build a replica of the current model? Will they beg the journal companies to lock up their papers in exchange for free work of writing, editing, reviewing and most importantly doing the entire research?

How the Current System Came into Existence

Before proposing an alternate system, let us review how we arrived at the current juncture. For a humorous take, you can check “How Biologist Got His Publishing System”, but here is what really happened.

Facilitated by the Industrial Revolution and the rise of industrial publishing, many now-famous journals came into existence in the mid-to-late 19th and early 20th centuries. In a 2013 post, I wrote about the first issue of Nature, published in 1869. The articles in that inaugural issue read more like blog posts reviewing important scientific discoveries than like original research reports.

Traditionally scientific societies took care of publishing. Scientists paid their yearly dues, and a significant part of the fund was used to maintain the publishing system. Also, the society members volunteered to review papers, because they felt responsible to maintain the quality of the journals. Members received the (paper) journals by mail and loved to read them.

Two developments fundamentally upended this system: (i) the arrival of the internet, and (ii) the massive increase in funding for the NIH.

Why is the System so Broken?

Yesterday I mentioned -

I was planning to buy a biology book published in the 1990s and wanted to check an old review on it before spending my money. The price of the thick book came down to below $20, but the one page review article still costs $32.99.

Why does a single short review from 1990 cost over $30? Here is why. Prior to arrival of the internet, the researchers transferred copyright to their articles to the journals to be published in physical media. This agreement was not meant to be for the internet. The journal companies are abusing that agreement to lock up old papers.

Thanks to the internet, a revolution took place in how we communicate with significantly lower costs, yet the journal companies applied the old method of pricing on the new media. The scientists went along, because the grants covered the funding. In other words, the internet revolution never arrived for scientific publishing. Even the open-access model is a minor tweak of the old system, because the journals charge an arm and a leg to allow you the privilege of open-accessing your paper.

Brainstorming for a New System

Just like the internet apps (e.g. Whatsapp) revolutionizing personal communications did not come from the phone companies, the changes in the publishing system will not come from the existing journals. Moreover, the government organizations like NIH will only issue edicts within the current system, and therefore the changes brought by them will be marginal.

I think we need a Whatsapp-like revolutionary change for publishing. The best way to think about it is to ask the question proposed earlier. If the current system disappears entirely, and the scientists gather to create a brand new publishing model from scratch, what will it look like?

I came across this idea from Mike Eisen.

I’ve said in many times and places what I think we should do. Have everything publishes as what we call now preprints and make that the only form of publication. It’s entirely author driven, and the infrastructure would be cheap and funded centrally so there would be no costs to publish or read. Then have peer review be layered on top of those already published works, with everything ranging from individual comments/reviews from readers to organized review by scientific societies to whatever else people invent to do it. The system of peer review we have today isn’t just expensive, it isn’t just bad because it denies lots of people access - it’s bad because it doesn’t do a good job at its job of assessing the validity and impact of science. We can do infinitely better for a lot less money if we cut journals out of the system entirely.

Do you have an alternate idea? Please post on twitter or email me (samanta@homolog.us), and I will be happy to add it here.