The Whiteboard - 2015-07-20 (page 1 of 2)

« first day (1781 days earlier) ← previous day next day → last day (3212 days later) »

12:00 AM

it's so close to being an extremely useful set of conventions I'd like to embrace, but then it turns out to have 2-3 different conventions for each little thing

@Snowman What' up?

I just went for a run. Need to take a shower, but I have a second.

@durron597 Needs to be cleaned up, but it's not off-topic.

@ThomasOwens It's not?

How else do we have questions about COCOMO?

Cost estimation in software engineering

== Methods == Methods for estimation in software engineering include this principles: Analysis Effort method COCOMO COCOMO™ II COSYSMO Evidence-based Scheduling Refinement of typical agile estimating techniques using minimal measurement and total time accounting. Function Point Analysis Parametric Estimating PRICE Systems Founders of Commercial Parametric models that estimates the scope, cost, effort and schedule for software projects. Proxy-based estimating (PROBE) (from the Personal Software Process) Program Evaluation and Review Technique (PERT) SEER-SEM Parametric Estimation of Effort, Schedule...

It need to be clear, via the tag wiki, that's what the tag is about. We may have 0 questions that are on-topic. But the tag shouldn't be blacklisted, because cost estimation is on-topic. It's part of project management.

@ThomasOwens We'll get there fast and then we'll take it slow?

user114359

@ThomasOwens Never mind, we noticed this user was making some... less than positive... posts. Looks like he has a network-wide suspension now.

user55340

12:02 AM

The difficulty that people have is "what does this cost" vs "how do I apply this cost estimation technique"

@MichaelT Yeah. The first isn't off-topic, but it is...too broad, I'd say.

user55340

6

Q: What should I keep in mind when writing custom software quotes?

I've been asked to give a quote to extend a piece of software i'm currently giving away for free for use by a company. They tell me they are using it in-house at the moment in conjunction with some "post processing" steps, which they would like to incorporate into the application proper. What sh...

billing cost-estimation

I would have closed it as too broad instead of primarily opinion based. But close enough.

I'd fix it if there was a meta complaint.

18

Q: Is it significantly costlier to fix a bug at the end of the project?

In a blog post by Andrew Hay, the following axiom was posited: It costs significantly more to fix a bug at the end of the project that it does to fix the same bug earlier in the project. However, this doesn't seem certain, especially after reading a blog post on Less Wrong, and the data I'v...

project-management debugging cost-estimation

technical debt does not need to be about money.

those two close reasons overlap a lot for some reason

12:04 AM

@durron597 That's on-topic.

user114359

@ThomasOwens it all started with this deleted question.

@ThomasOwens Why shouldn't that be retagged to technical-debt?

@durron597 I'm not sure if that's the right term.

that question seems like it's more about cost than technical debt

I think what he's getting at in that question is the cost of defect removal. A defect injected as a requirements defect may be more expensive to fix than a pure code mistake because the requirement influenced the design, the code, and the tests.

12:06 AM

he's not asking whether fixing a bug later or earlier results in more "debt" at release time

So if you find that you got a requirement wrong, you need to check and fix everything that came from that requirement. If you implemented a function wrong, then you only have the code and unit test to fix. The integration and acceptance tests are still likely right.

I'm not entirely convinced of that logic being sound. It's more likely to be true, but that's beside the point.

OK. I'm out. Need to take a shower. Be back around tomorrow.

Thanks, guys. Great help.

you're very welcome

user114359

12:27 AM

As of half an hour ago, I am a fanatic.

how did I get Fanatic before you?

user114359

@Ixrec Probably because I missed a day or two and it reset

1:03 AM

Maybe this question do not belong here (SOF)... I asked same question on Programmers (programmers.stackexchange.com/questions/290183/…) and can delete it from here if needed. Sorry. — Paruba 44 secs ago

user55340

click on the visited on the profile to see what days you visited.

user55340

user image

user55340

Green is days you've visited.

@Snowman I've got about a week to go on mine

Which is the command that defines a word in chat? or is that mod only?

user55340

2:12 AM

@durron597 define: $word

user55340

Verb: chat (third-person singular simple present chats, present participle chatting, simple past and past participle chatted)

To be engaged in informal conversation.
To talk more than a few words.
(transitive) To talk of; to discuss.
To exchange text or voice messages in real time through a computer network, as if having a face-to-face conversation.
chat

Noun: chat (countable and uncountable, plural chats)

(uncountable) Informal conversation.
A conversation to stop an argument or settle situations.
chat
chat (plural chats)
chat (plural chats)

(15 more not shown…)

user55340

That was define: chat

user114359

2:37 AM

Noun: snowman (plural snowmen)

A humanoid figure made with large snowballs stacked on each other. Human traits like a face and arms may be fashioned with sticks (arms), a carrot (nose), and stones or coal (eyes, mouth).

user114359

Wiktionary omitted the part about "snowmen are badass programmers"

2:49 AM

@Snowman If only they had 10 fingers to type with...

3:04 AM

I have 10 fingers of whiskey, and probably will lack the ability to type when I'm done!

user image

1 hour later…

4:34 AM

@durron597 too much whiskey: achieved. Typing status: nominal.

Myth: Busted.

3 hours later…

7:32 AM

I rarely ask here for something like that, but... can we please delete this garbage?

... In the name of the down-vote, the flag, and the vote to close. — Tim Post ♦ Jul 16 at 7:02

(original version was even worse)

4 hours later…

11:05 AM

@JamesThorpe I totally agree with you on the approach isn't ko best practice. I think delving into the SPA/Knockout/Data Binding best practices and design would be better suited on programmers — Anish Patel 31 secs ago

1 hour later…

12:23 PM

You're better off asking this question on Programmers — Jason Evans 39 secs ago

12:38 PM

5

Q: Why is a generic repository considered an anti-pattern?

it seems to me that a lot of specialised repository classes share similar characteristics, and it would make sense to have these classes implement an interface that outlines these characteristics, creating a generic repository to illustrate my point, say we have this code public class IEntity {...

Is it just me or are the comments on this question all completely wrong??

@MetaFight This comment isn't wrong.

without a C# background, I can't even work out what a "repository" is supposed to be; so what's wrong with all those comments exactly?

Can you provide a link to the sites that called this an anti-pattern? — Nimrand 24 mins ago

Of course it's not wrong. It's simply asking a question.

@MetaFight Yeah. That's a good comment.

>_> I know.

@Ixrec Repository is not a C# concept, it's a DDD concept.

not only are you wrapping a well known, tested repository (Entity Framework) inside a less feature rich interface, you are also artificially limiting the functionality that the consumers have, for little benefit. For example, why have an implementation of FindById that performs a where clause when db.Author.Find(someId) would do the same thing? — Claies 22 mins ago

this guy is essentially saying encapsulation is bad.

12:47 PM

unfortunately I have no idea what DDD is about either, maybe I'll just stay quiet

Domain-driven design

Domain-driven design (DDD) is an approach to software development for complex needs by connecting the implementation to an evolving model. The premise of domain-driven design is the following: Placing the project's primary focus on the core domain and domain logic. Basing complex designs on a model of the domain. Initiating a creative collaboration between technical and domain experts to iteratively refine a conceptual model that addresses particular domain problems. The term was coined by Eric Evans in his book of the same title. == Concepts == Context The setting in which a word or statement...

The main problem with that question is that the asker isn't actually using a "generic repository". He's using a generic "repository interface."

two very different things. The first is a leaky abstraction, the second is just a typing convenience.

Happy coffee day

Jul 14 at 14:00, by Ampt

@JimmyHoffa every day is coffee day

except for those of us who don't drink coffee

@Ixrec just because you're a heathenous non-celebrant, doesn't make it any less coffee day.

2

12:51 PM

I'm picking up Ixrec's slack by doubling my coffee consumption.

Ran out of creamer yesterday, so had my cup black this morning. Made me glad I'm always picky and spend the extra couple bucks for good coffee, it was still delicious.

It's already a 2 cup day.

I'm writing an SDD using my new approach. It's starting to gain traction.

@ThomasOwens "new approach" ?

the code-is-design approach?

@JimmyHoffa Kind of.

The thing was that design documents were never really well maintained. So rather than a design document being released early in the process, it's actually a deliverable at the same time as a software release.

You may need to issue the document before hand, or release if it's a customer deliverable, but the "final" document release of a design document for a given project is right with the deliverable media.

Throughout the requirements and all the way through writing code, the emphasis is on capturing design decisions somehow. It may be diagrams, text, whatever. But the document is just a snapshot at the end of what everything looks like.

The idea is to make it more useful for the next team. Assume that a whole bunch of new people are going to be working Version v+1, so tell them what they need to know, rather than propose the thing that you are building and put that under review.

1:09 PM

@ThomasOwens previously they were released early, or they are now?

@JimmyHoffa Yeah. They were released early and often incomplete. They were a "prove that you know enough about the problem to write code".

@ThomasOwens this is why good tracking tools that are not laden with endless fields, and allow a more freeform ad-hoc "Here's relevant information!" input approach are key to good documentation.

Meaning when someone new joined the project, they were thrown into code with no documentation about why decisions were made or where to begin with the project.

@ThomasOwens yeah, BDUF. Fail.

There are some decisions that need to be made early, but I think design documentation needs to be living. You just want to capture a snapshot minimally with every software release.

And that's what I'm trying to move toward. Have a document and update it with relevant information on a per-release cycle. The document is effectively a Word document wiki that gets PDF'd and released with a software release. It has slightly more structure than a wiki, though.

1:13 PM

The best living-documentation as you're describing I've ever worked with was at Employer^^^, we used TFS and it was expected/standard that whenever you did anything with any ticket/whatever, nobody ever filled out "fields" - they would just write their comments in the history tab - free form entry of whatever they felt was relevant with a timestamp. Done. Documentation would be compiled from these later if needed, but everyone in the org quickly learned just look their on a ticket to get started

@JimmyHoffa The problem with that is that we could have hundreds of tickets in a database, which are organized by project and not product.

every time I've ever worked somewhere with endless reams of fields in their tracking system to get that fine-grained tracking, just resulted in developers entering the minimal necessary required by management or whoever, and doing everything they could to avoid touching the system.

And the tickets are mostly defects or change requests.

@ThomasOwens true, but I'm simply referring to the concept and technique it showed of how to get organizational engagement on maintaining living documentation: Simple - free form - minimum level of entry

Yeah. The more free-form the better. Just enough consistency to be able to figure out what you're reading.

1:16 PM

@ThomasOwens yep, like I said that was the great thing about that system- free form but still provided to future readers consistency and a little extra data: Time stamp, user, and sequence automatically

The more you can derive automatically from the lowest-level-of-input, the better. For instance what if alongside timestamp each history entry identified the current sprint? Anywho, that's my 2 cents on living documentation. For what you're doing, wiki may be the best approach (that's what I wrote my req doc in just the other day)

We're going to be migrating to Jira (and that whole suite), so there may be better options going forward.

But unfortunately, config management is still very much document-centric and expects Word or PDF things to be controlled.

brb cup 3

My last company had JIRA - total nightmare. THey customized the hell out of it and it never made any sense - every 2-3 months they'd create a new JIRA project because the last one wasn't meeting their needs since they kept modifying shit and nobody ever used it. New company uses JIRA- so far it's pretty damn simple. Has a free form comments area just like the TFS history.

Moving states isn't "fill out these 15 required fields which only management understands" - it's, drag the ticket from one swimlane to the next.

The biggest thing about all of those tools is - more often than not - do not customize them. Learn their out of the box suggested process, use them in the lightest way possible, and you'll generally have good engagement and something useful.

Hell if you do customize them, do it to remove requirements to their user rather than adding it. Developers positive engagement with those tools is the absolute best way to derive value from any of them, and no matter how great the tools are, their value will be nil if your devs aren't engaged with it - no matter how much you force them to use it. Use and engagement are totally different things.

Yeah.

I've been on a "kill the process steps that don't add value or map to compliance" rampage recently.

Once (if) my edits get approved, would this be a good candidate for migration here?

5

Q: Why is a generic repository considered an anti-pattern?

it seems to me that a lot of specialised repository classes share similar characteristics, and it would make sense to have these classes implement an interface that outlines these characteristics, creating a generic repository to illustrate my point, say we have this code public class IEntity {...

That's one major step, the other one is usability. You want devs engaged with reporting tools, sit down with them and see how they use them and what their pain points are just like you would with any users you were supporting. One of the constant annoyances at my last place was their network and security infrastructure was totally busted so JIRA was incapable of integrating and remembering credentials. Every time anyone wanted to use JIRA, they had to log in again.

1:30 PM

@MetaFight sure, why not

Getting rid of small shit like that so the tools fit in the flow of their normal work better can be just as important as removing the process steps they make them hate the tools.

@ThomasOwens what kind of software are you writing/documenting? a lot of this sounds interesting but probably not relevant to my workflow

It's good to know what devs are actually aware of too - fields about "product", "project", "release", "iteration" - these are more often then not, things devs aren't aware of or thinking about, making them fill them out will cause them to avoid the tool. Those are fields management put for themselves to know, but usually devs aren't in the meetings where the littany of projects or products or releases are necessarily talked about. Depends on the org anyway.

@Ixrec he's an engineer, he works on bridges, not software.

ok, what kind of bridge?

@Ixrec I make software for this. I've worked on the test and diagnostics, mission planning, image processing, command and control...

I've helped test the avionics side, too. But not so much development there.

1:42 PM

@JimmyHoffa sooooooo agree, the only problem I have with Jira (other than the slowness, which is our company's fault) is that it imposes a heck of a lot of completely unnecessary restrictions that essentially mean a lot of state changes have to be done as 2-3 state change sinstead of just 1

my favorite is you can't add a story point estimate to anything in the current sprint; that is just so dumb

user114359

@Ixrec You can use security to lock that down. Making corrections is important, but estimating in the current sprint is madness.

@Ixrec don't know if that's default or customized? I haven't used JIRA a lot here yet so I may have just not hit the pain points yet.

I believe it's the Agile plugin; I think it doesn't have sprints at all by default

perhaps the more I use it I'll realize even the defauly configuration is irritating as hell.

user114359

I have actually seen it where a story had 50 points because someone added a zero on accident, and we caught it during the sprint.

1:45 PM

afaik we only use the one plugin, though I havent worked with Jira elsewhere so hard to tell

user114359

Greenhopper is the Scrum plugin

@Snowman if you follow the golden ideal of agile every waking moment, yes this doesn't happen

is my edit here too drastic? stackoverflow.com/review/suggested-edits/8862203

in reality, our scheduling meeting has a fixed time and sometimes deciding what goes into the sprint is more important than getting everything estimated

hm, could've sworn our ops referred to it as the "Agile plugin"

user114359

@MetaFight after a quick read-through it appears you did improve the question, but your changes were very drastic

1:48 PM

@JimmyHoffa I usually drink my coffee black - if only because I don't get shit from my on-a-diet gf. I'm practically on a diet by proxy

the other reason is that a lot of our tickets go unestimated because they're either little bugs (and deciding which bugs are showstoppers is more important than deciding which are 2 or 3 when we have so few left in total), or because they're vague requirements that need UX/biz clarification before estimation makes any sense

I have one edit approval and one edit rejection. In my opinion, the only real problem with the original question was the incorrect usage of the term "Generic Repository." Maybe I did go a bit too far by changing his example from an interface to an abstract base class. I couldn't resist refactoring though!

user114359

@Ixrec It sounds like while you may be doing Agile development, you are not adhering strictly to Scrum.

it's likely that we aren't

user114359

Not that it is a bad thing, I have said before and I will say it again that I like the Agile principles but dislike the Scrum implementation of those principles.

1:50 PM

the distinctions at the level of Agile v Scrum and TDD v BDD are the point where I generally stop caring about the theory because there are more practical real-world issues to deal with first

like the fact that some large chunks of the core of our codebase are still effectively untestable

but for the sake of discussion, how exactly do you define "scrum"?

user114359

@MetaFight Personally if the question is so broken it needs that much editing, then so is my give-a-damn. If the question needs a complete rewrite, I do not want to risk changing it too much.

user114359

@Ixrec Scrum are the processes, that is where stuff like "sprints" and "Scrum master" come from.

@MetaFight my knee-jerk response is "refactoring goes in the answer, not a question edit"

@Snowman in theory we do sprints and we supposedly had a scrum master (before he moved to Germany), is there a specific definition you had in mind? and which bit of what I said struck you as violating that definition?

@Ampt

'atom-workspace':
  'alt-x': 'command-palette:toggle'
  'ctrl-x 0': 'pane:close'
  'ctrl-x 1': 'pane:close-other-items'
  'ctrl-x 2': 'pane:split-down'
  'ctrl-x 3': 'pane:split-right'
  'ctrl-x b': 'fuzzy-finder:toggle-buffer-finder'
  'ctrl-x ctrl-c': 'application:quit'
  'ctrl-x ctrl-f': 'fuzzy-finder:toggle-file-finder'
  'ctrl-x ctrl-j': 'tree-view:reveal-active-file'
  'ctrl-x k': 'core:close'
  'ctrl-x o': 'window:focus-next-pane'

^-- Ok, maybe I will be able to work with Atom... :)

back to Jira for the moemnt

a better example probably would be that the actions you're allowed to take on a ticket vary a lot depending on what screen you're on; sometimes Withdraw is a single click, other times it's a combo box inside a popup inside a dropdown

1:57 PM

@Ampt I've gone back and forth. At my last job they had starbucks beans which were good enough to have a cup on their dime with creamer, but I can't drink that crap black. Here they have costco beans which are awful so I'm supplying my own coffee again, which means good stuff like Hiline or Peets which I can enjoy black.

half the time when we're in backlog grooming meetings there's a few instances of us "backseat Jiraing" as our team lead tries to update a ticket correctly

@Ixrec yes - this is another scenarios where tools are in your way rather than helping. Years ago it was pretty common in industry for meetings to involve someone assigned as note taker, sometimes it would rotate but there was always someone assigned the task. I miss those days because there was no tools bullshit getting in the way of just working out whatever the meeting was for. After the meeting notes would get posted, and any tasks like updating systems or documents would happen then.

admittedly, we have the tools in part because what we need to discuss has no hope of fitting on a notepad or whiteboard

For whatever reason everybody thinks they have to be interacting with technology 24/7 and multi-tasking and bullshit which is crap - so these classical single-taskers like note-taker don't exist. Everyone's expected to wear multiple hats, no matter how much they end up tripping over their own feet.

tbh if we could just get our internal Jira instance to run faster it wouldn't really be in the way at all

actually we do have a designated note taker when we demo to business; that works very well

user114359

2:04 PM

@Ixrec one of the invariants of Scrum is you cannot work on something that is not estimated. You can't even put it in the current sprint because you have no idea if you have the velocity for it.

I think we're getting away with it because, at least right now, almost all unestimated tickets are fairly small (especially the "bugs reported yesterday" variety)

user114359

Bugs don't always need stories anyway, especially if they are small and reported on a story in the current sprint.

@gnat Do you think a VLQ flag would have been marked as helpful here in July 2015

what if they're not related to any stories in the current sprint?

I think right now almost everything in the sprint is bug fixes

user114359

@Ixrec then you need a story.

2:08 PM

cool

@Ixrec good point.

user114359

2:20 PM

39

Q: Which 2015 technologies were correctly predicted by Back to the Future II?

This year is not only the 30th anniversary of the Back to the Future franchise as a whole, but is also the year in which Back to the Future II is set. Which technologies that exist in 2015 (at least as prototypes) were correctly predicted by Back to the Future II ? To clarify, I am looking ...

back-to-the-future back-to-the-future-ii

I'm surprised Wad's only gotten 45 upvotes for his answer to that so far

Hendo Hoverboard: The World's First Hoverboard

Oh.

Hendo Hoverboard: BUSTED

I thought it was real but only worked on magnetic surfaces or something

user114359

I am not all the way through reading the answers but I think there is another kind of meta-answer. The maker movement which has existed for many years even if we didn't have that name for it, has a lot of extra tools now. Arduino, 3D printers, etc. we could make all sorts of cool stuff.

91 degrees. Feels like 99 degrees. 10:26am.

:|

2:27 PM

"maker movement"?

user114359

That won't give us flying cars or "real" hoverboards, but pretty much everything else I can think of from that movie either exists or is within reach, only not existing because there is no need for it

user114359

The maker culture is a contemporary culture or subculture representing a technology-based extension of DIY culture. Typical interests enjoyed by the maker culture include engineering-oriented pursuits such as electronics, robotics, 3-D printing, and the use of CNC tools, as well as more traditional activities such as metalworking, woodworking, and traditional arts and crafts. The subculture stresses a cut-and-paste approach to standardized hobbyist technologies, and encourages cookbook re-use of designs published on websites and maker-oriented publications. There is a strong focus on using and...

@Ixrec Yeah, that's what the debunk video basically said.

seems like it, I was just skimming it

user114359

I didn't watch the video but yes, the hoverboards I have read about all have limitations.

2:28 PM

making a rideable board out of magnets is still nothing to sneeze at though

@Ampt ok - I'm sincerely beginning to like atom. Their initial goal (or my understanding of it?) spoke to me: emacs with extensibility in a language I can make sense out of - elisp feels too dated. I'm able to edit my keymappings and add and handle packages in it very easily, and with one package it's behaving like emacs with tiling and buffer navigation...

The biggest thing I like about emacs is the ability to tile my buffers and navigate them all from the keyboard. Toss in REPL buffers and it's great; Atom has the same tiling and navigation built in - you just need to map the keys, and the keyboard mapping file is easy as can be to edit being JSON (well, CSON), and REPL buffers are one package away

user114359

I also like how the Cubs beat the Miami Marlins in the World Series. The Marlins were not a team (yet) in '91 or whenever the movie was made, and both teams are in the NL anyway.

though I just learned that killing a buffer with unsaved changes pops up a dialog box that asks to save changes with yes as the default. C-x k enter did not kill and get rid of my changes as expected, but killed and saved them...now to find a default config.cson because mine's corrupt... :)

personally, I find the "don't configure it" rule applies to editors as much as it does tools like Jira

which is probably why I've never had much of an opinion on editors beyond "vim's cool, I use it sometimes, haven't tried the others"

@Ixrec tools are used by a single sole person; JIRA is used by a bunch of them which is why don't configure it is a good rule - or rather, don't configure it worse which is what 99% of people do... Editors on the other hand, making those form to your mind like a glove definitely enhances productivity... getting used to using various hot keys in visual studio or any editor I've worked with has significantly improved my ability to not have my train of thought interrupted

2:41 PM

there's enough about my dev environment I have no control over that I can't really apply that argument; sometimes I have to switch machines, or use the in-house tooling, or fight with configuration management systems where even the defaults don't work much less the advanced stuff my colleagues tried

if I do a one-man project on one single machine for fun, then I'd probably configure a bit more because I'd actually have real control over it

@Ixrec yes well, these are more significant issues than just local usage..

sounds like you've got a bit of a mess; but you like C++ so of course you're taking it on the chin with a smile in place heh

bizarrely, this is just as true of our Javascript as our C++

but yes it's mostly our fault

I just mean non-C++ folk haven't been brutalized into the strange "The pain just doesn't hurt! I love my overlords! Make the tools harder, master!" stockholm syndrome displayed by C++ devs

;P

oh no I do not have that

I definitely complain about it

on the C++ side, let's just say we have an entire in-house tool whose sole purpose is to simulate large-scale linking to work out what dependencies are missing from the linkline; we have to run it pretty much every time we work on a C++ microservice

and that's actually an improvement; previously we had a different tool that tried to do something similar and didn't even work half the time

on the Javascript side, for some reason the team for our in-house framework felt the need to develop their own custom IDE instead of giving us command line tools (or any other way of escaping from said IDE); if the thing actually worked reliably I wouldn't mind so much, but...

for now the battle I've chosen is to get them to fix the unit test runner in said IDE, because that's by far the easiest high-value thing to fix that doesn't go against any strange value differences between our departments

@JimmyHoffa False, configure the fuck out of jira just because you can! Build it from source with different jars, configure a dozen active directories, and have a rsync script to backup the assets folder every night!

/me weeps

3:00 PM

so, totally random discussion-y question

28

Q: How appropriate is the interview question: "Show me a piece of code you like"?

I recently interviewed a candidate at our company and asked him Show me a piece of code you like, yours or something you snatched from the internet The candidate was not really able to properly answer that, just fumbled around aimlessly and unsure of what to make of this kind of response...

interviewing software-industry

the responses to this seem to be in the vein of "no, I couldn't come up with something off the top of my head, I'd have to go look for it"

but I'm honestly not sure I would be able to find any code I actively "liked", it feels to me like good code is just code that you never have to think about, much like a good UI is one you never notice because it doesn't get in the way of your actual task

@Ixrec If I were asked that I'd come up with something gimmicky

but that's not really something you want to advertise as code you like

3:29 PM

Monday Morning Coffee Jam

gahh I still haven't made my second cup yet!

pflech.

@gnat @GlenH7 @Ixrec Do we want this? stackoverflow.com/questions/31519332/…

Enough of this shit, I need to go clean the french press and get cracking.

user41796

@durron597 Not really

@durron597 it's pretty "meh" tbh

user41796

3:37 PM

It's analogous to "discuss this blog"

user41796

Bold claims need to be backed by more than a paragraph. Otherwise it's just flamebait

we get quite a few "someone said X, why did he say that?" questions which inevitably have to be closed because we have no idea what X was thinking when he said it

if the question is "are RDBMS's suitable for distributed stuff" then the answer is probably "um, duh?" so I dunno if we'd want that

@durron597 quite likely to be declined (that is, if flagged answer is accepted with solid score). I keep an eye at MSO discussions and it looks like this attitude didn't go away. It is as if mods learned to and now prefer to suffer through some pain of meta beating in cases when flagger complains

He's not making the claim. If the question were "Are RDBMS suitable for distributed database systems?" would it be okay?

given some explanation of a reason why he suspected why they might be unsuitable, I would think so

user41796

3:39 PM

@durron597 Begs the question of "what or who says they aren't suitable?"

but there doesn't appear to be such a reason

@durron597 too strong a scent of Discuss this ${blog}

@gnat VLQ is treated differently than NAA

@gnat That could be edited out.

@durron597 VLQ? means that answer has score 0 or lower? in this case, I'd say chance for it to get through moderator today are close to 90%. Low score is a different game

@gnat Oh, that was a high scoring accepted answer? mehhhh

user41796

3:41 PM

You could certainly argue that traditional RDBMS have to jump hoops in order to support a distributed approach. But it's been done, and it remains to be seen that any other approach has a superior alternative to resolving that hoop jumping. The underlying issues behind distributed transactions are thorny and are an orthogonal issue to the DB management style

@durron597 if edited out, I wouldn't mind having it here

@durron597 per my recollection yes, it was high score accepted. You have 10K, you can check vote split over there (accept mark disappears on deletion, that would be hard to verify)

@gnat What about this version? stackoverflow.com/questions/31519332/…

user41796

Still really big

I don't think that addresses the core issue of "duh? why do you think they wouldn't be?"

user41796

And the supporting quote appears to be whinging more about consultants than underlying technology issues

3:45 PM

I'm looking at the original article he's quoting from now

I'm voting to close this question as off-topic because it's not about specific code; it seems more suitable for programmers.stackexchange.com. — chepner 9 secs ago

I'm not seeing any assertion over there of an inherent or fundamental reason why RDBMS must be slower, only an appeal to personal experience that NoSQL apparently handles that better

eg, "I know from experience that as a database grows in size or the number of users multiplies, many RDBMS-based sites suffer serious performance issues."

What's the link to view timeline on an answer again?

user41796

Looks like the blog author is a consultant who plays the stereotypical, tired tripe of "consultants are bad"

is it just posts/idnumber/timeline ?

3:46 PM

and "If an organization is facing such issues, then it should consider NoSQL technologies, as many of them were designed specifically to address these scale (horizontal scaling or scale-out using commodity servers) and performance issues."

user41796

@durron597 yes

user41796

@Ixrec 'cause he wouldn't be biased at all by advising for mongo

nope

I think the only part of that section of the article that appears to give a "first principles" reason for it would be this:

"NoSQL databases, in general, avoid RDBMS functions like multi-table joins that can be the cause of high latency. In the new world of big data, NoSQL offers choices of strict to relaxed consistency that need to be looked at on a case-by-case basis."

@durron597 still somewhat slippery but since I can't point what exactly is wrong - it's maybe just a paranoia. I'd give it a chance over here

which to me at least, seems like a special case of "it's fast because it has fewer features"

I'm not sure if that's really a criticism of RDBMSs per se

user41796

3:48 PM

@gnat It's too broad as in "If a whole book could be written about it..."

also ^that

@gnat If I'm reading this correctly it was zero score, accepted.

user41796

@Ixrec "It's fast because it'll let you corrupt your data."

heck, you could argue it's a technology recommendation question

@GlenH7 consistency is a feature =)

@durron597 what was vote split? I ask because meta effect could drop it

3:49 PM

Looks like gnat and I vote for migration, glen votes against. Ixrec?

user41796

I once had to work with IVR software that had explicitly chosen "fast" over "accurate"

I'm definitely against

user41796

The vitriol I can spew about that product... <sigh>

@gnat I got that number from timeline, before meta effect

so now we're 2/2 on migration, bleh

if I didn't sound like it, that's only because I was focusing on trying to find the kernel of a good question in it, but thus far I have failed

user41796

3:49 PM

@durron597 - It would be more interesting to start from the ground up with that question

user41796

There could be an interesting and constructive question, but at that point, make it your own and don't worry about the salvage + migrate

I don't care enough to write my own.

@durron597 I think this is the real sign of whether it's a good enough question or not

@Ixrec No, I think it's a good question, I just do very little with large scale databases and I have zero experience with NoSQL

I should not be the person authoring this question.

@GlenH7 too broad, maybe. I re-read, and still have no firm feeling. Guess I wouldn't VtC / would skip it in the queue. I hate too broad because it's often so complicated (save for few clearly defined cases like where-to-start / pros-and-cons etc)

user41796

3:51 PM

@durron597 And apparently neither should they. :-D

@GlenH7 Ha!

well, at the very least we can all agree that the original form of the question was essentially guaranteed to get closed here as "discuss this blog"

user41796

yep, I'd agree with that

I'd try suggesting a more on-topic version myself but I still haven't worked out wtf NoSQL actually is

user41796

@Ixrec NoSQL really just means non-relational database

3:54 PM

Duga comment incoming...

user41796

but "non-relational DB" is a mouthfull

@GlenH7 that's kinda the problem, it's defined as being not something else, that doesn't tell me what it is

user114359

amazon.com/NoSQL-Distilled-Emerging-Polyglot-Persistence/dp/…

in the philosophical jargon, we have a negative definition but not a positive one

user41796

and what non-relational means is that the traditional rules of normalization and whatnot within the data don't make as much sense for certain data types

3:55 PM

You may want to consider posting this question on Programmers.SE. I have not flagged this question for migration because, in it's current form, it is too broad. What is the problem that you're having that you are trying to solve with a database solution? I recommend that you self delete this question, and post a new question on Programmers with more details about your particular problem without so much focus on what you read in a book. — durron597 59 secs ago

user41796

time series data is perfect fit for a non-relational DB

user114359

@Ixrec good book, I bought a copy. Doesn't go into too much detail but does cover everything you need to know about NoSQL to get started.

user41796

Or said another way, not all data is a good fit for a relational DB

user41796

graphs can get shoehorned into an RDBMS but aren't a great fit

user41796

some non-relational DBs handle those domains better

3:56 PM

@Snowman thanks, I'll see if I feel like reading a technical book this week

g2g for a bit

user41796

If you ignore the hype of "NoSQL for everything!!!" then you realize it's just another tool for particular domains

@GlenH7 I picked NoSQL first for basketball night

user41796

with doing statistical analysis on teams and whatnot?

@GlenH7 you should see its jump-shot

user41796

Ah, gotcha

user41796

3:58 PM

He brings an awesome and fanatical fan base with him to every game

user41796

Of course, they don't always understand the sport that's being played. But they're always loud and supportive.

@GlenH7 I know, right? Cheerleaders at a pick-up game

user41796

It's completely awesome. Even if some of their cheers don't make sense or are outright offensive

Someone reviewed this as "looks ok" ughhhhhhh

user114359

@durron597 If I had enough SO rep I would review that with a shovel

4:03 PM

@Snowman You do have 125 rep...

user114359

@durron597 I am a bit shy of 3k

@Snowman I meant you can downvote it.

user114359

@durron597 Meh, I need more rep on SO, not less.

@Snowman You don't get -1 for downvoting questions, only answers. That's true network-wide

user41796

question downvotes are free

user114359

4:10 PM

I didn't realize that. I never paid attention I guess.

0

Q: Can we give people an easy way to analyze previous reviews?

I think that the current review system offers an awesome opportunity to learn about how people do in their reviews via crowd sourcing. For example, I'd like to be able to see how often my reviews disagree with those of others: I was inspired by looking at this review task In case it gets delete...

feature-request review statistics

4:46 PM

@this: I was being polite. (This site can't afford to be NSFW.) Someone who sets standard error to fully buffered is subverting the reasonable expectations of programmers. Such people should not be allowed to release code for other people to use. — Jonathan Leffler 12 secs ago

well...shit. Here I was trying to get JIRA working in emacs this morning, when I figured I'd try Atom- got it all emacsified, but just went and found out, no jira plugin. Bla. So guess it's back to emacs.

user114359

Emacs integrates with Jira? Wow, its reputation really is warranted: Emacs is its own operating system! There is nothing it cannot do!

5:20 PM

@Snowman It's not as good at basketball as NoSQL

It does bring its own cheerleaders just the same, though.

user114359

@durron597 Emacs cheerleaders are gross. Any description I can think of is NSFW.

user41796

@Snowman but they'll pound the living vim out of you....

6:09 PM

This question belongs on programmers.stackexchange.com — PA. 18 secs ago

I'm voting to close this question as off-topic because this question belongs on Programmers.SE, except it would be closed as a duplicate of this question — durron597 42 secs ago

So this is interesting... in the last 90 days, there have been zero rejected migrations from SO to programmers.

out of 21 questions.

\o/

Meanwhile, Code Review has had 60 migrations, with 16% rejection.

They've had more migrations than TeX and GIS combined.

user55340

6:29 PM

@Snowman it doesn't integrate with lldb because it makes RMS confused about the license and goals.

6:41 PM

Welp, I've officially been lied to by comcast. Where do I sign up for the TShirts?

user55340

@Ampt send mail to nobody@comcast.com. They've assured me this is the support desk and will send the shirt via drone carrier.

> I got scammed by Comcast and all I got was this lousy t-shirt

Don't worry, I'm now talking to their ridiculously overpriced business internet

I was told I'm getting 150 down/ 25 up with a 600GB cap

I'm actually getting 105D/25U with a 300GB cap

« first day (1781 days earlier) ← previous day next day → last day (3212 days later) »

Transcript for

Jul '1520

The Whiteboard

General Discussion for softwareengineering.stackexchange.com P...

coffee-day horror-stories monads my-code-is-compiling scotch the-other-monitor

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile