8 ms·

Ask HN: Examples of bad open-source code to learn what to avoid?

66 points by theSage 7y ago

What are some of the bad examples of code that you have seen? Something you would want to avoid? I'm looking for examples which fall along the lines of "fail to see the forest for the trees".

codr7 7y ago
It doesn't work, energy follows thought and it makes no sense to focus it on what you want to avoid. Take or leave. That being said, the worst I ever saw was the in-house business nonsense I was paid to deal with as a Java consultant. The worst code isn't open source from my experience, subjecting it to public scrutiny would mean suicide for the companies involved.
- SmellyGeekBoy 7y ago
  This matches my experience. Open source projects tend to get the worst parts fixed. It's in-house applications, usually written in VB / Delphi / Java, that have been supported and added to over the past 20 years where the true horrors lie. As you say, I doubt many businesses would be willing to put this kind of code out there.
neya 7y ago
I feel like this could lead to very opinionated, non-constructive comments/flame, but if were to give you an example, I'd suggest taking a look at Wordpress eco-system. While Wordpress core's codebase has improved significantly over the years, some of the plugins haven't. The top pick goes to WooCommerce, although an open source E-commerce solution on top of Wordpress, it has some terrible decisions under the hood. The top pick would go to mixing presentational logic with business logic. For example, to render a table, instead of exposing an array of objects to allow the developer to loop through it as he/she sees fit, WooCommerce will force you to use a PHP function that renders a table for you and there's actually no way to modify the presentation logic if you wanted to. It's a really fundamental programming paradigm that even top open source companies fail to adhere to. Again, I'm not saying this to attack them or the maintainers behind the code, just my opinion of why I think it's bad quality code while respecting the fact that developers still do take time and effort for us to enjoy something with freedom and zero cost.
- sleavey 7y ago
  Agreed. The core code is actually pretty well written considering the concrete backwards compatibility standards enforced by the devs, but most plugins are a total mess (and I'm saying that as a plugin author). I guess it is mainly due to PHP's low bar for entry and PHP's historical (but arguably not now) lack of enforcement of good programming practices. There is potential for change though. There is a project backed by the core devs (can't remember the name) which will score plugins by their code standard and show the score in the plugin directory.
  - allenskd 7y ago
    I feel it's not that PHP is low bar but rather widely accessible out of the box on almost all web hosting services... you don't have to set up a lot of stuff before running your web application. I'd argue the only reason PHP still remains at the top is because no one has made X language to work out of the box with cpanel installations and run with little to no modification to run, lets say, django apps or RoR apps. Maybe if Python or any other language spent more time on improving that type of accessibility in the realm of web applications we could see PHP fragmenting in the later years...
    - combatentropy 7y ago
      I think the main reason I still prefer PHP after doing web development for over a decade is not its syntax but its set-up. Picture me, often the only programmer, supporting several-dozen business web apps. Each app is in its own directory. To add a route, I just make a new PHP file. This eases mingling other, static, files in that same directory: CSS, JS, HTML, PDFs, documents, spreadsheets, images, and videos that the users wanted to supplement the app with. Maintaining a list of routes in some other file would complicate things, especially with the number and complexity of the apps that we have. (User-uploaded files are stored elsewhere.) That is really a feature of CGI, and FastCGI, not PHP. But I like PHP more than Perl, and I'm not sure how well the FastCGI frameworks for other languages are written. For example, I often dream of using JavaScript in FastCGI. But then again I am also spoiled by the extent of PHP's standard library. It's mainly a bunch of functions in the global namespace, but an alphabetic namespace is vast, and there have been just a handful of times that I wanted to make a function whose name was already taken. PHP is also stable. I've never had a problem with a script bringing down other parts of the website or even other parts of the same app (except that one time when a coworker wrote a PHP script that built SQL dynamically and did it in a bad way, and used up all the RAM on the server). PHP has been fast for me, even back in 2005. Then again I don't use third-party frameworks. I have written a few helper libraries, to reduce code repetition.
yamann 7y ago
https://github.com/mholt/caddy https://github.com/mholt/caddy not only a mediocre code, but the guy behind it received lots of money from Mozilla as an innocent promising open source project author, then he made it as a paid product.
AnaniasAnanas 7y ago
Here you go https://github.com/progwml6/Natura/blob/1.7.10/src/main/java/mods/natura/common/NContent.java https://github.com/progwml6/Natura/blob/1.7.10/src/main/java... Tip: If you ever end up in a situation where you have to copy-paste code with minor changes then there is something that you are doing wrong. In this case using arrays and loops would be a much better solution.
- thrower123 7y ago
  Good god, my eyes, they burn.
- thih9 7y ago
  But: avoid premature optimisation, it could lead to overly complex abstractions. Simple, clear structures with little repetition are ideal. Copy pasted data is painful to work with but still relatively easy to maintain and fix. Overly complex and unintuitive abstractions are the most difficult to maintain and improve.
  - ilaksh 7y ago
    But it's very repetitive code and uses identifiers like "var3" and "var4".
- allenskd 7y ago
  The function addShapedRecipeFirst (List recipeList, ItemStack itemstack, Object... objArray) is really doing a number on me... gonna be fun not touching that code for a few months and when you come back you don't even know what the heck you are doing
- pjc50 7y ago
  That looks nasty, but surprisingly hard to fix with loops because everything is of different types. If I were fixing it I'd look to some kind of code generation solution, even if just a hacky python script parsing a CSV. (addShapedRecipe is just begging to have ASCII art as its canonical form)
  - 7y ago
slezyr 7y ago
See Syobon Action it's source code as bad as the game itself. https://github.com/angelXwind/OpenSyobonAction/blob/master/main.cpp https://github.com/angelXwind/OpenSyobonAction/blob/master/m...
twhitmore 7y ago
Libraries can be useful despite imperfections, and poor design decisions can occur in overall good libraries. So we can't judge too harshly. Having said that: 1) iText PDF library used to have some fairly poor & duplicated code. Column layout was a highlight. Also strange ideas overemphasizing subclasses, eg. for paragraph styles. (Correct approach: use values rather than types.) 2) Tomcat webserver back around 2007 used to have some amazing 'clustering' code to deploy your webapp across multiple servers. But it lacked proper knowledge & hence control of what it was doing. IIRC there was no clear master, and a server couldn't tell what had been started on it versus what had been replicated since a peer was seen to be running it. Effect: replication would be additive only, contexts would just replicate everywhere uncontrolled, and there was no good way to stop/ undeploy an app across the cluster.
spion 7y ago
How about examples of open source code to learn whats really, really good, together with why it was designed that way? Seems like that would be way more useful. Or examples of projects that did things one way, but later refactored, and why they refactored.
- allenskd 7y ago
  I think someone starting, or intermediate, or perhaps advanced won't be able to tell what's good or bad... even I have problems sometimes identifying what would be the best approach of implementing a good design and could end up implementing a bad one easily. And refactoring as almost always a must... unless you are stuck in a legacy support project where you are just hacking fixes away despite the glaring flaws and the client doesn't want to spend any more money on improving things.
- convolvatron 7y ago
  i dont think you really get very deep into it by reading code without working with it. sure, there are surface syntactic niceties one can bikeshed. the real meat of the matter comes when you are trying to make a change. is the structure robust? is there convenient tooling that helps you do what you need to do? does the system require extensive boilerplate to do simple things? does the system come crashing down in some unrelated area when you are trying to make simple changes? it may be surprising, but large old codebases usually have huge hunks that serve no real purpose whatsoever except to glue together two pieces that would be much happier talking directly to one another. I really wish as a community we could abandon the 70s business notion that software is a concrete artifact that one invests in and sells. its a really poor model. software is a process. code that is not being maintained is largely just dead. as developers we should be evaluating software as a living thing that responds to its environment...not as a shrink wrapped item we unbox and review on youtube.
sam_lowry_ 7y ago
Jgit is amazingly bad for a piece of software built on top of well thought out data structures if git. Some if its flaws could be attributed to Java IO design, though.
ddebernardy 7y ago
OpenSSL (before HeartBleed) springs to mind. https://news.ycombinator.com/item?id=7640378 https://news.ycombinator.com/item?id=7640378
- blattimwind 7y ago
  OpenSSL code and docs are still a huge mess.
- shoo 7y ago
  banter and patches from OpenBSD here: https://opensslrampage.org/ https://opensslrampage.org/
blattimwind 7y ago
Drupal 6/7 would be an example for "widely used [at the time] but pretty bad". I don't know how many of the issues across all layers were addressed in later versions. OpenSSL is still an excellent example for very messy code where even maintainers / frequent contributors regularly get lost. Also a good example for designing many bad APIs and poor docs. libsodium is a good counterexample, although the internal structuring of the code base is a bit atypical, it is logical and consistent. (It does have some API idiosyncrasies which cater specifically to dynamic bindings, like providing a constant always as a #define/macro but also as an exported function; and it has a bit of an issue where you have both legacy APIs and newer APIs, but the docs are pretty clear on which is which). BorgBackup is an example of how you don't want to mix C and Python code, and also contains various bits that only 1-2 people on the planet really bothered to understand, besides demonstrating other issues of organically grown code bases.
pjc50 7y ago
> fall along the lines of "fail to see the forest for the trees" I think I'm going to need an example to understand this? Having said that, one of the most informative programming books I've ever read was C Traps And Pitfalls. Flags common easily-made errors and explains them, which in turn fixes misconceptions about the language. I feel most languages could do with one.
jasode 7y ago
An example of bad code that always stuck with me was the flawed CDDB disc id hash algorithm.[0] I was reminded of that short-sighted decision every time I ripped a bunch of CDs and saw how importing song titles was not automatic because a dozen different discs had the same hash ids which resulted in collisions[1]. It ended up creating needless friction for millions that depended on that discid. What's sad is I'm not even sure if one can extract any useful "lessons learned" from it! The programmer that wrote it was not an amateur script kiddie; he had a computer science degree from Uni California. Apparently, he didn't realize he was writing a flawed hash algorithm as he wrote it. One could say that hash algorithms should be "peer reviewed". Well, he got unsolicited peer review that pointed how his homegrown hashing computation was flawed but he ignored the suggestion to improve it. [0] >Ti Kan wanted to use a hash. He could have chosen something like CRC32, which would have given him a 32 bit number, yielding 4 billion unique IDs. Instead he wrote his own hash. [...] Ti Kan was made aware (not by me) of this problem back in 1994, and given a script to convert this format into a CRC32-based format, but he rejected it because the deployed base was too big. At that point it was probably in the high dozens. -- excerpt from http://quimby.gnus.org/circus/notes/cddb.html http://quimby.gnus.org/circus/notes/cddb.html [1] https://forums.macrumors.com/attachments/multiple-matches-jpg.17857/ https://forums.macrumors.com/attachments/multiple-matches-jp... [2] wiki: https://en.wikipedia.org/wiki/CDDB#How_CDDB_works https://en.wikipedia.org/wiki/CDDB#How_CDDB_works
- the8472 7y ago
  That advice is wrong too, a 32bit number would have been insufficient due to the birthday problem.
  - jasode 7y ago
    I don't think any reasonable person would expect zero collisions. The weakness of CRC32 for even distribution of hash values was well-known. (CRC32's goal was a fast "checksum" instead of strong cryptographic hash.) The point was CDDB's non-invented-here home-grown hash algorithm was worse than CRC32. He didn't extract the maximum entropy from the discs' metadata of song times to minimize future collisions.
  - 7y ago
RickJWagner 7y ago
Just a side note-- this is the beauty of Open Source. If there is some bad code (we all write it), it can be improved with a little help from Open Source "friends". All of us are stronger than any one of us. Long live Open Source!
arthev 7y ago
While usually focusing on smaller snippets, thedailywtf.com is a site about (mostly) code wtfs. A fair number of management wtfs too, though.
bradknowles 7y ago
There are an infinite variety of things that you don’t want. If you focus all your energy on those things, then you won’t have anything left to do the positive things you do want. You have to turn that equation around. Even if all you know right now is the negative thing you don’t want, you have to figure out how to reframe that into the positive thing you do want. On,y then can you make positive progress towards that thing you want — and by the way, you will naturally avoid the things you don’t want by focusing on the things you do want. Sure, examples of bad stuff can be instructive, but only so far as it helps you further clarify the good stuff you’re actually trying to achieve.
superpermutat0r 7y ago
I've always found Calibre to be a huge mess.
ben_bai 7y ago
Always a crowd-pleaser: OpenBSD true.c vs. GNU's true.c http://cvsweb.openbsd.org/cgi-bin/cvsweb/~checkout~/src/usr.bin/true/true.c?rev=1.1&content-type=text/plain http://cvsweb.openbsd.org/cgi-bin/cvsweb/~checkout~/src/usr.... https://github.com/coreutils/coreutils/blob/master/src/true.c https://github.com/coreutils/coreutils/blob/master/src/true....
- k4ch0w 7y ago
  Simple is better imo, openbsd wins.
kissgyorgy 7y ago
That's a really bad idea. You need to have a good counterexample, otherwise it's just wasting time at the best case. A lot of people learn from really bad codebases and picking up the same style which is terrible. You should look at GOOD codebases instead!
- theSage 7y ago
  I did some work on html2text when I started off and while the little parts did make sense, the whole library was confusing for me. I couldn't change anything without breaking tests. On the other hand I've been tinkering with curio for a while now and it's a fresh breath compared to that. My trouble is that I still don't understand what makes the html2text thing "bad". What particular thing there caused me to not like working with it? I'm trying to understand that. I've been book hunting + figuring out if it's something that I did not know which would have made the code a lot easier to work with (stuff got a lot easier after I cleaned up my set theory understanding) - https://github.com/dabeaz/curio https://github.com/dabeaz/curio - https://github.com/Alir3z4/html2text https://github.com/Alir3z4/html2text
otras 7y ago
If you’re interested in a case of unnecessary optimization and effort, the infamous left-pad npm library has been refactored to only add to the string O(log(n)) times. It is short but not sweet. https://github.com/left-pad/left-pad#readme https://github.com/left-pad/left-pad#readme
- ksaj 7y ago
  I think it needs more // comments. Hilarious.
Sir_Cmpwn 7y ago
Here's some old code of mine: https://github.com/vatt849/LibMinecraft/blob/master/LibMinecraft/Client/MultiplayerClient.cs https://github.com/vatt849/LibMinecraft/blob/master/LibMinec... The whole library is a trip if you want to read a bunch of bad C#. Highlights: - Generated documentation - Giant switch/case instead of a more organized dispatch map - Large swaths of commented code instead of using version control - try...catch statements that just eat the errors - Inconsistent code style - This thing: https://github.com/vatt849/LibMinecraft/blob/master/LibMinecraft/Client/MultiplayerClient.cs#L603-L621 https://github.com/vatt849/LibMinecraft/blob/master/LibMinec... I've written something similar from scratch since, which I'm still not entirely satisfied with, but is much better for reference: https://github.com/ddevault/TrueCraft https://github.com/ddevault/TrueCraft The client-side networking code lives here: https://github.com/ddevault/TrueCraft/blob/master/TrueCraft.Core/Networking/PacketReader.cs https://github.com/ddevault/TrueCraft/blob/master/TrueCraft.... https://github.com/ddevault/TrueCraft/blob/master/TrueCraft.Client/MultiplayerClient.cs https://github.com/ddevault/TrueCraft/blob/master/TrueCraft.... https://github.com/ddevault/TrueCraft/tree/master/TrueCraft.Client/Handlers https://github.com/ddevault/TrueCraft/tree/master/TrueCraft.... Notable improvements: - Handwritten docs only where necessary - Uses a stream implementation for decoding this particular wire format - Has a different and better abstraction for reading packets out Still has bad error handling though.
frostburg 7y ago
Praat: https://github.com/praat/praat https://github.com/praat/praat I'm not sure that this is a good way to learn anything, however.
peterwwillis 7y ago
Anything related to OpenStack, but particularly jenkins-job-builder is rather horrible.
type0 7y ago
"How to Make Mistakes in Python" - https://www.oreilly.com/library/view/how-to-make/9781492048275/ https://www.oreilly.com/library/view/how-to-make/97814920482...
ksaj 7y ago
Here is something atrocious I wrote in Lisp. You actually can make Lisp ugly! Who knew? https://github.com/ksaj/Capitalize.Lisp https://github.com/ksaj/Capitalize.Lisp
jstarfish 7y ago
The source for Terraria is notoriously terrible. https://github.com/TheVamp/Terraria-Source-Code https://github.com/TheVamp/Terraria-Source-Code It is another example of how even inelegant code full of hardcoded values can be successful.
- czr 7y ago
  (Note that this is decompiled source – the original probably at least has comments here and there)
craftoman 7y ago
Many JavaScript libraries like Fastify (Node.js) for example. You always get a nice & clean API but if you look under the hood you would be amazed at how much spaghetti code can be written in a project.
jxub 7y ago
Many, if not most OSS packages which are released by academics or universities.

66 points by theSage 7y ago

What are some of the bad examples of code that you have seen? Something you would want to avoid? I'm looking for examples which fall along the lines of "fail to see the forest for the trees".

codr7 7y ago

It doesn't work, energy follows thought and it makes no sense to focus it on what you want to avoid. Take or leave. That being said, the worst I ever saw was the in-house business nonsense I was paid to deal with as a Java consultant. The worst code isn't open source from my experience, subjecting it to public scrutiny would mean suicide for the companies involved.

SmellyGeekBoy 7y ago
This matches my experience. Open source projects tend to get the worst parts fixed. It's in-house applications, usually written in VB / Delphi / Java, that have been supported and added to over the past 20 years where the true horrors lie. As you say, I doubt many businesses would be willing to put this kind of code out there.

neya 7y ago

I feel like this could lead to very opinionated, non-constructive comments/flame, but if were to give you an example, I'd suggest taking a look at Wordpress eco-system. While Wordpress core's codebase has improved significantly over the years, some of the plugins haven't. The top pick goes to WooCommerce, although an open source E-commerce solution on top of Wordpress, it has some terrible decisions under the hood. The top pick would go to mixing presentational logic with business logic. For example, to render a table, instead of exposing an array of objects to allow the developer to loop through it as he/she sees fit, WooCommerce will force you to use a PHP function that renders a table for you and there's actually no way to modify the presentation logic if you wanted to. It's a really fundamental programming paradigm that even top open source companies fail to adhere to. Again, I'm not saying this to attack them or the maintainers behind the code, just my opinion of why I think it's bad quality code while respecting the fact that developers still do take time and effort for us to enjoy something with freedom and zero cost.

sleavey 7y ago
Agreed. The core code is actually pretty well written considering the concrete backwards compatibility standards enforced by the devs, but most plugins are a total mess (and I'm saying that as a plugin author). I guess it is mainly due to PHP's low bar for entry and PHP's historical (but arguably not now) lack of enforcement of good programming practices. There is potential for change though. There is a project backed by the core devs (can't remember the name) which will score plugins by their code standard and show the score in the plugin directory.
- allenskd 7y ago
  I feel it's not that PHP is low bar but rather widely accessible out of the box on almost all web hosting services... you don't have to set up a lot of stuff before running your web application. I'd argue the only reason PHP still remains at the top is because no one has made X language to work out of the box with cpanel installations and run with little to no modification to run, lets say, django apps or RoR apps. Maybe if Python or any other language spent more time on improving that type of accessibility in the realm of web applications we could see PHP fragmenting in the later years...
  - combatentropy 7y ago
    I think the main reason I still prefer PHP after doing web development for over a decade is not its syntax but its set-up. Picture me, often the only programmer, supporting several-dozen business web apps. Each app is in its own directory. To add a route, I just make a new PHP file. This eases mingling other, static, files in that same directory: CSS, JS, HTML, PDFs, documents, spreadsheets, images, and videos that the users wanted to supplement the app with. Maintaining a list of routes in some other file would complicate things, especially with the number and complexity of the apps that we have. (User-uploaded files are stored elsewhere.) That is really a feature of CGI, and FastCGI, not PHP. But I like PHP more than Perl, and I'm not sure how well the FastCGI frameworks for other languages are written. For example, I often dream of using JavaScript in FastCGI. But then again I am also spoiled by the extent of PHP's standard library. It's mainly a bunch of functions in the global namespace, but an alphabetic namespace is vast, and there have been just a handful of times that I wanted to make a function whose name was already taken. PHP is also stable. I've never had a problem with a script bringing down other parts of the website or even other parts of the same app (except that one time when a coworker wrote a PHP script that built SQL dynamically and did it in a bad way, and used up all the RAM on the server). PHP has been fast for me, even back in 2005. Then again I don't use third-party frameworks. I have written a few helper libraries, to reduce code repetition.

yamann 7y ago

https://github.com/mholt/caddy https://github.com/mholt/caddy not only a mediocre code, but the guy behind it received lots of money from Mozilla as an innocent promising open source project author, then he made it as a paid product.

AnaniasAnanas 7y ago

Here you go https://github.com/progwml6/Natura/blob/1.7.10/src/main/java/mods/natura/common/NContent.java https://github.com/progwml6/Natura/blob/1.7.10/src/main/java... Tip: If you ever end up in a situation where you have to copy-paste code with minor changes then there is something that you are doing wrong. In this case using arrays and loops would be a much better solution.

thrower123 7y ago
Good god, my eyes, they burn.
thih9 7y ago
But: avoid premature optimisation, it could lead to overly complex abstractions. Simple, clear structures with little repetition are ideal. Copy pasted data is painful to work with but still relatively easy to maintain and fix. Overly complex and unintuitive abstractions are the most difficult to maintain and improve.
- ilaksh 7y ago
  But it's very repetitive code and uses identifiers like "var3" and "var4".
allenskd 7y ago
The function addShapedRecipeFirst (List recipeList, ItemStack itemstack, Object... objArray) is really doing a number on me... gonna be fun not touching that code for a few months and when you come back you don't even know what the heck you are doing
pjc50 7y ago
That looks nasty, but surprisingly hard to fix with loops because everything is of different types. If I were fixing it I'd look to some kind of code generation solution, even if just a hacky python script parsing a CSV. (addShapedRecipe is just begging to have ASCII art as its canonical form)
- 7y ago

slezyr 7y ago

See Syobon Action it's source code as bad as the game itself. https://github.com/angelXwind/OpenSyobonAction/blob/master/main.cpp https://github.com/angelXwind/OpenSyobonAction/blob/master/m...

twhitmore 7y ago

Libraries can be useful despite imperfections, and poor design decisions can occur in overall good libraries. So we can't judge too harshly. Having said that: 1) iText PDF library used to have some fairly poor & duplicated code. Column layout was a highlight. Also strange ideas overemphasizing subclasses, eg. for paragraph styles. (Correct approach: use values rather than types.) 2) Tomcat webserver back around 2007 used to have some amazing 'clustering' code to deploy your webapp across multiple servers. But it lacked proper knowledge & hence control of what it was doing. IIRC there was no clear master, and a server couldn't tell what had been started on it versus what had been replicated since a peer was seen to be running it. Effect: replication would be additive only, contexts would just replicate everywhere uncontrolled, and there was no good way to stop/ undeploy an app across the cluster.

spion 7y ago

How about examples of open source code to learn whats really, really good, together with why it was designed that way? Seems like that would be way more useful. Or examples of projects that did things one way, but later refactored, and why they refactored.

allenskd 7y ago
I think someone starting, or intermediate, or perhaps advanced won't be able to tell what's good or bad... even I have problems sometimes identifying what would be the best approach of implementing a good design and could end up implementing a bad one easily. And refactoring as almost always a must... unless you are stuck in a legacy support project where you are just hacking fixes away despite the glaring flaws and the client doesn't want to spend any more money on improving things.
convolvatron 7y ago
i dont think you really get very deep into it by reading code without working with it. sure, there are surface syntactic niceties one can bikeshed. the real meat of the matter comes when you are trying to make a change. is the structure robust? is there convenient tooling that helps you do what you need to do? does the system require extensive boilerplate to do simple things? does the system come crashing down in some unrelated area when you are trying to make simple changes? it may be surprising, but large old codebases usually have huge hunks that serve no real purpose whatsoever except to glue together two pieces that would be much happier talking directly to one another. I really wish as a community we could abandon the 70s business notion that software is a concrete artifact that one invests in and sells. its a really poor model. software is a process. code that is not being maintained is largely just dead. as developers we should be evaluating software as a living thing that responds to its environment...not as a shrink wrapped item we unbox and review on youtube.

sam_lowry_ 7y ago

Jgit is amazingly bad for a piece of software built on top of well thought out data structures if git. Some if its flaws could be attributed to Java IO design, though.

ddebernardy 7y ago

OpenSSL (before HeartBleed) springs to mind. https://news.ycombinator.com/item?id=7640378 https://news.ycombinator.com/item?id=7640378

blattimwind 7y ago
OpenSSL code and docs are still a huge mess.
shoo 7y ago
banter and patches from OpenBSD here: https://opensslrampage.org/ https://opensslrampage.org/

blattimwind 7y ago

Drupal 6/7 would be an example for "widely used [at the time] but pretty bad". I don't know how many of the issues across all layers were addressed in later versions. OpenSSL is still an excellent example for very messy code where even maintainers / frequent contributors regularly get lost. Also a good example for designing many bad APIs and poor docs. libsodium is a good counterexample, although the internal structuring of the code base is a bit atypical, it is logical and consistent. (It does have some API idiosyncrasies which cater specifically to dynamic bindings, like providing a constant always as a #define/macro but also as an exported function; and it has a bit of an issue where you have both legacy APIs and newer APIs, but the docs are pretty clear on which is which). BorgBackup is an example of how you don't want to mix C and Python code, and also contains various bits that only 1-2 people on the planet really bothered to understand, besides demonstrating other issues of organically grown code bases.

pjc50 7y ago

> fall along the lines of "fail to see the forest for the trees" I think I'm going to need an example to understand this? Having said that, one of the most informative programming books I've ever read was C Traps And Pitfalls. Flags common easily-made errors and explains them, which in turn fixes misconceptions about the language. I feel most languages could do with one.

jasode 7y ago

An example of bad code that always stuck with me was the flawed CDDB disc id hash algorithm.[0] I was reminded of that short-sighted decision every time I ripped a bunch of CDs and saw how importing song titles was not automatic because a dozen different discs had the same hash ids which resulted in collisions[1]. It ended up creating needless friction for millions that depended on that discid. What's sad is I'm not even sure if one can extract any useful "lessons learned" from it! The programmer that wrote it was not an amateur script kiddie; he had a computer science degree from Uni California. Apparently, he didn't realize he was writing a flawed hash algorithm as he wrote it. One could say that hash algorithms should be "peer reviewed". Well, he got unsolicited peer review that pointed how his homegrown hashing computation was flawed but he ignored the suggestion to improve it. [0] >Ti Kan wanted to use a hash. He could have chosen something like CRC32, which would have given him a 32 bit number, yielding 4 billion unique IDs. Instead he wrote his own hash. [...] Ti Kan was made aware (not by me) of this problem back in 1994, and given a script to convert this format into a CRC32-based format, but he rejected it because the deployed base was too big. At that point it was probably in the high dozens. -- excerpt from http://quimby.gnus.org/circus/notes/cddb.html http://quimby.gnus.org/circus/notes/cddb.html [1] https://forums.macrumors.com/attachments/multiple-matches-jpg.17857/ https://forums.macrumors.com/attachments/multiple-matches-jp... [2] wiki: https://en.wikipedia.org/wiki/CDDB#How_CDDB_works https://en.wikipedia.org/wiki/CDDB#How_CDDB_works

the8472 7y ago
That advice is wrong too, a 32bit number would have been insufficient due to the birthday problem.
- jasode 7y ago
  I don't think any reasonable person would expect zero collisions. The weakness of CRC32 for even distribution of hash values was well-known. (CRC32's goal was a fast "checksum" instead of strong cryptographic hash.) The point was CDDB's non-invented-here home-grown hash algorithm was worse than CRC32. He didn't extract the maximum entropy from the discs' metadata of song times to minimize future collisions.
- 7y ago

RickJWagner 7y ago

Just a side note-- this is the beauty of Open Source. If there is some bad code (we all write it), it can be improved with a little help from Open Source "friends". All of us are stronger than any one of us. Long live Open Source!

arthev 7y ago

While usually focusing on smaller snippets, thedailywtf.com is a site about (mostly) code wtfs. A fair number of management wtfs too, though.

bradknowles 7y ago

There are an infinite variety of things that you don’t want. If you focus all your energy on those things, then you won’t have anything left to do the positive things you do want. You have to turn that equation around. Even if all you know right now is the negative thing you don’t want, you have to figure out how to reframe that into the positive thing you do want. On,y then can you make positive progress towards that thing you want — and by the way, you will naturally avoid the things you don’t want by focusing on the things you do want. Sure, examples of bad stuff can be instructive, but only so far as it helps you further clarify the good stuff you’re actually trying to achieve.

superpermutat0r 7y ago

I've always found Calibre to be a huge mess.

ben_bai 7y ago

Always a crowd-pleaser: OpenBSD true.c vs. GNU's true.c http://cvsweb.openbsd.org/cgi-bin/cvsweb/~checkout~/src/usr.bin/true/true.c?rev=1.1&content-type=text/plain http://cvsweb.openbsd.org/cgi-bin/cvsweb/~checkout~/src/usr.... https://github.com/coreutils/coreutils/blob/master/src/true.c https://github.com/coreutils/coreutils/blob/master/src/true....

k4ch0w 7y ago
Simple is better imo, openbsd wins.

kissgyorgy 7y ago

That's a really bad idea. You need to have a good counterexample, otherwise it's just wasting time at the best case. A lot of people learn from really bad codebases and picking up the same style which is terrible. You should look at GOOD codebases instead!

theSage 7y ago
I did some work on html2text when I started off and while the little parts did make sense, the whole library was confusing for me. I couldn't change anything without breaking tests. On the other hand I've been tinkering with curio for a while now and it's a fresh breath compared to that. My trouble is that I still don't understand what makes the html2text thing "bad". What particular thing there caused me to not like working with it? I'm trying to understand that. I've been book hunting + figuring out if it's something that I did not know which would have made the code a lot easier to work with (stuff got a lot easier after I cleaned up my set theory understanding) - https://github.com/dabeaz/curio https://github.com/dabeaz/curio - https://github.com/Alir3z4/html2text https://github.com/Alir3z4/html2text

otras 7y ago

If you’re interested in a case of unnecessary optimization and effort, the infamous left-pad npm library has been refactored to only add to the string O(log(n)) times. It is short but not sweet. https://github.com/left-pad/left-pad#readme https://github.com/left-pad/left-pad#readme

ksaj 7y ago
I think it needs more // comments. Hilarious.

Sir_Cmpwn 7y ago

Here's some old code of mine: https://github.com/vatt849/LibMinecraft/blob/master/LibMinecraft/Client/MultiplayerClient.cs https://github.com/vatt849/LibMinecraft/blob/master/LibMinec... The whole library is a trip if you want to read a bunch of bad C#. Highlights: - Generated documentation - Giant switch/case instead of a more organized dispatch map - Large swaths of commented code instead of using version control - try...catch statements that just eat the errors - Inconsistent code style - This thing: https://github.com/vatt849/LibMinecraft/blob/master/LibMinecraft/Client/MultiplayerClient.cs#L603-L621 https://github.com/vatt849/LibMinecraft/blob/master/LibMinec... I've written something similar from scratch since, which I'm still not entirely satisfied with, but is much better for reference: https://github.com/ddevault/TrueCraft https://github.com/ddevault/TrueCraft The client-side networking code lives here: https://github.com/ddevault/TrueCraft/blob/master/TrueCraft.Core/Networking/PacketReader.cs https://github.com/ddevault/TrueCraft/blob/master/TrueCraft.... https://github.com/ddevault/TrueCraft/blob/master/TrueCraft.Client/MultiplayerClient.cs https://github.com/ddevault/TrueCraft/blob/master/TrueCraft.... https://github.com/ddevault/TrueCraft/tree/master/TrueCraft.Client/Handlers https://github.com/ddevault/TrueCraft/tree/master/TrueCraft.... Notable improvements: - Handwritten docs only where necessary - Uses a stream implementation for decoding this particular wire format - Has a different and better abstraction for reading packets out Still has bad error handling though.

frostburg 7y ago

Praat: https://github.com/praat/praat https://github.com/praat/praat I'm not sure that this is a good way to learn anything, however.

peterwwillis 7y ago

Anything related to OpenStack, but particularly jenkins-job-builder is rather horrible.

type0 7y ago

"How to Make Mistakes in Python" - https://www.oreilly.com/library/view/how-to-make/9781492048275/ https://www.oreilly.com/library/view/how-to-make/97814920482...

ksaj 7y ago

Here is something atrocious I wrote in Lisp. You actually can make Lisp ugly! Who knew? https://github.com/ksaj/Capitalize.Lisp https://github.com/ksaj/Capitalize.Lisp

jstarfish 7y ago

The source for Terraria is notoriously terrible. https://github.com/TheVamp/Terraria-Source-Code https://github.com/TheVamp/Terraria-Source-Code It is another example of how even inelegant code full of hardcoded values can be successful.

czr 7y ago
(Note that this is decompiled source – the original probably at least has comments here and there)

craftoman 7y ago

Many JavaScript libraries like Fastify (Node.js) for example. You always get a nice & clean API but if you look under the hood you would be amazed at how much spaghetti code can be written in a project.

jxub 7y ago

Many, if not most OSS packages which are released by academics or universities.