Hacker News new | past | comments | ask | show | jobs | submit login
ASCIIFlow Infinity (asciiflow.com)
368 points by billpg on Jan 2, 2018 | hide | past | web | favorite | 56 comments

There's also Monodraw, ASCII drawing with a native MacOS GUI and more advanced tools.


Monodraw is incredible, and vastly more useful than ASCIIFlow just by dint of the fact that objects remain objects after drawn - you can drag them around and the graph remains intact.

ASCIIFlow does keep models of drawn boxes, lines, and arrows. If you use the "resize/move boxes and lines" tool, you can adjust drawn objects.

Ah, so it does. Not especially well, but it does. My mistake!

Author here - it does but it doesn't, which is why it doesn't always work so well. It never actually keeps track of the shapes in a seperate data model - the text itself is the only data model. When you resize, it retraces out the lines that need editing. I did this as the expected use case is for diagrams in code, where diagrams may be copy pasted in and out of the tool and I'd lose any underlying models that I was working with. For fun: https://github.com/lewish/asciiflow2/blob/master/js-lib/draw...

Grateful user here. I've relied on ASCIIFlow for all the diagrams in my Internet drafts since IETF 88. Truly an essential tool in my arsenal.

Wow - I just love this! Thanks very much for the info - feels like a new pair of shoes for 2018 (coming from 15 years of omnigraffle)...

I've been using Monodraw for awhile now and it's been great!

See also previous discussions: https://news.ycombinator.com/from?site=asciiflow.com

Also, this looks like it is the repository for ASCIIFlow and you can run it locally:


Reminds me of JavE (http://jave.de/).

EDIT: What would be extra awesome would be a mode where it saves SVG. And/or a tool that renders SVG as ASCII art.

EDIT: Also, JavE does much more than asciiflow, as it even allows you to create ASCII movies.

I like this. Unfortunately it's not as semantic as I expected:

     |     XX          |
     |      XX         |
     |       +XXX------+
     |       |   XXXXX
     +-------+       XXX

     If I try to resize
     a rectangle that's
     crossed by a free-form
     curve, it stops being
     a rectangle.
(Now that I think, I could also use this as a feature).

This is explained by a comment from the author (bootywizard) above:

Author here - it does but it doesn't, which is why it doesn't always work so well. It never actually keeps track of the shapes in a seperate data model - the text itself is the only data model. When you resize, it retraces out the lines that need editing. I did this as the expected use case is for diagrams in code, where diagrams may be copy pasted in and out of the tool and I'd lose any underlying models that I was working with. For fun: https://github.com/lewish/asciiflow2/blob/master/js-lib/draw....

This could be fixed by being more expansive/creative with choice of glyphs, so that two intersecting lines get represented by a glyph that uniquely corresponds to the 2 glyphs that overlap at that position (possibly also encoding z-depth stack in the glyph choice)

It makes sense. That's a good design decision.

Cool! I can use it to make some ASCII arts for my console applications.

However, please change the icon of Freeform Drawing Tool. Because currently that icon make it looks like a Curve Drawing Tool in the first glance.

There is also a bug when you press a key during freeform drawing on Firefox Nightly. It doesn't properly set the active character. Works fine on Chrome though.

the freeform drawing is not that good, this works so much better http://www.jave.de/ - example trace I did with that https://pastebin.com/raw/8gedcN63

As a Linux sysadmin, this is incredible flowchart software and I really enjoy using it. Is there a diagram software for console I could use thats just as easy I wonder?

The downside is it has very limited sex appeal for *.managers. Microsoft Visio has turned diagrams into blinding tapestries of overpriced clipart, fancy arrows, and logos. Worst of all, its an enabler for the cardinal sin of diagrams: overlapping lines.

Check out PlantUML - there's a nice demo here: https://www.planttext.com/

Plenty of editors support it, both console (vim/emacs) and GUI-based (VSCode, Sublime).

Checkout yEd (Google it) built ontology of graphml, it makes restructuring diagrams a doddle.

Very cool! It would be awesome if it had a mode that used the unicode box-drawing characters too [0]. It can get a little tedious copy and pasting them and drawing things manually.

- [0] https://en.wikipedia.org/wiki/Box-drawing_character

On similar lines, I very recently discovered the combination of emacs + Artist Mode + ditaa for creating diagrams, and the experience has been pretty good so far too.

Oh yeah, this combo is actually one of the better modes in Emac-land. Whenever I see yet another ASCII art link posted here, I just gaggle cos Artist mode is that good.

Like many emacs things, there more than one. There's also picture-mode. I haven't used either very much, but my sense is that artist-mode is more mouse-oriented and possibly supports more sophisticated drawings but picture-mode may work better in terminals.

Pretty cool, but the text tool really needs some figlet-like functionality with fonts and outlines!

Great tool. I used it to create a state machine diagram - one that I then included in the inline documentation of my Elixir code.

Clever. I once had to diagram a state machine for a PR to an OSS program since the code is a bit inscrutable if you don't understand the purpose. This would have been quite useful in that regard.

Great execution!

I understand the impetus to keep things simple. But PNG image export would be a tremendous boon to me. Should be trivial as you are just drawing text to the canvas, no?

As a workaround you can just take a screenshot (although that could get cumbersome with large diagrams).

quick emacsen, let's augment artist-mode to match it (especially the topology friendly resize)

Wow, that smart resize tool is awesome, and makes this a good tool for roughing out dungeon maps. :)

I used to use asciiflow for some of my network sketching, and it can be enjoyable to play with. That said, I have been thinking about the subject of network maps for quite some time now as a senior sysadmin, and I no longer use it for a couple of reasons.

First, I don't like the idea of being tied to a webservice for my map drawing tools. I would really like the source code so I can self-host. I looked around the new updated asciiflow website and didn't see any links to source so correct me if I'm wrong.

Second, and most importantly, as a senior sysadmin over the years I have seen hundreds of network maps, the vast majority of them being in visio due to it having market dominance, and there is a problem that neither visio or it's competitors has addressed, and that's automated network mapping with versioning so you can diff and walk back in time. It's the thing even the best automappers still don't do right because they tend to get the topology wrong (such as one of my favorite automappers like librenms's).

This is why one of the utilities I have been working on is a bash-script cobbling together nwdiag (http://blockdiag.com/en/nwdiag/index.html), a subset of the blockdiag, with nmap and (h/f)ping scanners as the generating input and cronified daily versionable text output with checksums, with bonus things like new kea-dhcp lease triggering rescan via the kea restful api.

I think the future of network maps needs to be both automated and text based to enable admins and executives the ability to fully visualize the network at any given state in time, with the ability to walk the cat back with confidence in the accuracy of the map you are looking at(I would like to integrate gpg as well).

It's not quite ready for release, (sitting in my github projects folder waiting for push), but I lazily named it anmap for automated-network-mapper.

So right now how I use it personally is I have an emacs-org mode file that is cronified to output daily latex pdf reports that calls the bash script and generates the output map which is then listed inside the rest of the org-mode systems report which does other stuff like give me go-no-go on system usage via stuff like lynis, rkhunter, ckrootkit, and ossec.

tldr - Basicallly the network map industry is extremely ripe for competition and all these web-based network mapping tools are still not addressing a core needs usecase for sysadmins.

I haven't tried it, but could it be this one? https://github.com/lewish/asciiflow2

That looks like it to me, though for some reason I didn't expect it to be java. Regardless, thanks so much for the link! Huge bonus that it's gplv3 for me, I will def be playing with this. Did you find it from the website and I just missed it or did you have to search for it?

A few weeks ago I was interested in this project, too, and looking around I came across this link. The author's profile also points to the AsciiFlow website, so I assumed that would be it. I think you might have misread the language; it's actually written in javascript :-)

I'm going to blame my phones autocorrect, but it was mostly because I hadn't had coffee yet.

Awesome application! I used it in the past to draw ASCII diagrams for my blog posts (e.g. [0]) as I was looking for an alternative way to display "images". Still needs some tweaking for mobile usage, but on desktop it's just great because of the removed loading times for images.

- [0] https://www.robinwieruch.de/learn-react-before-using-redux/

I've never seen that "Can you help us?" Google donation tab before.

Is that new? It looks like Google requires you to register as a non-profit to use it. Am I looking at the wrong thing?


There's also my favorite, Asciio https://metacpan.org/pod/App::Asciio

It allows saving the object format for later modification or export as pure ascii text, as needed. And it works on all platforms that support Perl (a lot).

Love seeing more tools for good old ASCII art, but it seems ANSI art has become a forgotten relic of the past...

This is great. I've made a few diagrams with it today, and they've been easy to do and came out just right.

I do wish I could toggle the background and view the diagram on blank white; having the grid visible makes some things look more obvious than without, and that would be useful to know ahead of time.

Another project we worked on using machine learning (decision tree) to transform an input image or video frame into printable ASCII characters:


In Firefox, arrows that point to boxes have the head set to '^' regardless of their actual direction.

So an arrow to a box might look like ----^

I just tried it with Firefox (latest). I didn't have that issue. When I moved the line the > became ^ until I dragged it back into place.

It is a great tool. I use it to create simple GUI mockup, then copy-paste it to github/gitlab issue.

I'm working on a prototype that's basically made for what you're trying to do. A text command based UX tool. You can create both text-renderings and image-renderings: https://medium.com/proof-of-concept/creating-a-true-ux-tool-...

I wrote a number of articles on it, and soon will open up an online version of the prototype. Love to get your feedback.

Similar, but for MOTD on linux/unix boxes

this is so cool, is there any plan to make it work in a terminal?

Emacs has a built-in mode for doing this - picture-mode

Did you get this to work on a modern emacs ?

Wow! Works very well on my iPhone, too! Great work.


Why is there a separate tool for creating text? It would be better to just click to place the cursor and start typing.

Also this works on mobile.

So that you can paste?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact