neomem blog

well those clever spammers got into the forum somehow. i must admit i haven’t even looked at in a long time - keeping up an open source project is a lot of work if it gets even somewhat popular - there’s a lot of feedback to keep up with. which is why i’ve never advertised or promoted neomem anywhere, except way back when it was shareware, though even that was pretty minimal. so sorry if i’ve neglected the forum. i’m trying to get back into dealing with everything now though. i’d like to overhaul this site also, and move it all into wordpress, especially if they have a good forum plugin.

that’s also why i’ve been avoiding moving the source code to sourceforge - i sort of dread getting a barrage of emails about how to do x in the code, why is x done this way? how should x be done? etc. but if this thing is ever going to have a decent life i’d need to do that, and be prepared to deal with a lot of correspondence.

so, i don’t know. maybe it would be easier to start the open source project with the new version, which would be plugin based and more modular, therefore less obscure for people to contribute to the code. i’m also thinking about using the eclipse rich client, which is just SO nice. that would mean using java (or maybe groovy?), but that would be nice because it could also use db4o for the backend (a real oo database).

Tags: No Tags

i’m working on fixing an annoying bug - the richedit control (used in the Text View and in searching the entire file) sometimes bombs when rtf text is loaded into it, eg when you do a search, or clicking on an item. at least it does on some people’s systems - i haven’t been able to replicate it, even when someone sent me a file that was bombing! it might be some difference in richedit dll versions (people might have different versions on their systems). i should have a prerelease version of 1.2c fairly soon that will either fix the problem or at least provide some more information on what’s going on. it will also have some more export formats - i’ll try and squeeze xml export into it before the final release.

Tags: No Tags

well, neomem got put on the back burner for a while, but i’m trying to get back to it. it’s sort of tricky in that it can tend to be extremely consuming. but maybe if i limit the time spent on it i’ll actually get more done. ie avoid burnout!

most likely features to be worked on next: xml export, dynamic/search folders, better filtering options, and import csv files. along with some code refactoring.

i’ll also be working on the next version - mostly just the data backend for now. i think i’ve figured out how i’d like it to handle single/multiple values, which had been a bit of a roadblock for a while. well, we’ll see if it works, after a bit of prototyping…

p.s. i switched to moderated comments due to a flood of spam. i’ll try installing one of those spam blocker plugins though.

Tags: No Tags

yeah, i think the best thing i could be doing at the moment is work on refactoring the code, organizing it and making it easier to understand for other developers (and myself, too!). sourceforge did get their subversion hosting working this month, so i’ll be able to host the code there, and i’ll probably migrate the forum and feature database over there. i’d much rather have neomemRails working to host it all, but i just don’t know how long that would take…

there are a lot of features to add to neomem, and there’s no way i can do it all myself, so making it easier on other developers would be pretty important…

and i’m slowly getting to old emails and forum posts…

Tags: No Tags

Version 1.2 is now available here

New features include:

• Export entire file or selected object (recursively) to csv (this has limitations, as far as being imported to other programs - Excel has a 32k character per cell limit and Access a 64k character per row limit, which makes this feature less useful than one might hope. But *someday* there will be an export to XML feature…)
• Uses an ini file instead of the registry for storing options
• Added File, Folder, Email, and Website property types, to replace the Hyperlinks type (which is still available, though deprecated)
• Added Strikethrough format
• Insert date/time from Edit menu or F5 key
• Use Alt+, and Alt+. for back/forwards
• Use Alt+Backspace for undo, in addition to Ctrl+Z
• Various bug fixes, including the one of printing extra blank pages

I’ll send out an announcement email tomorrow - there’s an unofficial mailing list that I need to make official somehow, ie with a signup form on the website and everything.

Tags: No Tags

as i’m putting the finishing touches on the next version, i was thinking, this code really needs some refactoring. it was the first c++ and object-oriented program i wrote, and it shows… i still don’t quite know the best way to do things in c++ - eg things that are so simple in ruby are extremely complex (and brittle) in c++. which is why i don’t like it. but maybe if i spent some time cleaning up the code, modularizing things, and learning the right way to do things in c++, it would be a good investment, because if i’m going to make this a real open source project, it really should have cleaner structure. eg adding a garbage cleanup library would be nice, though the ones i’ve seen don’t work with mfc…

also, it would be an opportunity to either get it ready for, or transition to, wxWidgets (or Qt?) (away from mfc), and also away from visual c++ to the eclipse cdt, which i’d really love to be using. and this would also make it cross-platform (theoretically - not sure how good wxWidgets is at that). and maybe doing some redesign as i go, using test-driven development, a concept which i really like, and wish i had known about when i started neomem back in 1999.

anyway, i actually enjoy doing refactoring, and making things more organized. the only thing is, it doesn’t get me closer to having the type or program that i actually want! neomem has lots of limitations, though maybe i could be throwing new features in as i went…

the biggest limitation to me, at this point, is not being able to just instantly create a new object, not caring about where it goes in a hierarchy, or if the class and property structures are set up yet. that’s what is always slowing me down, and why i still have two boxes full of index cards… so maybe just use ctrl+n for a new object, bring up a dialog with a grid at the top for properties, and a big scratch space at the bottom, and if you just want to throw a bunch of text in there for now, that’s fine. you could always go back and refactor the object later and partition the text out to properties.

the rails project kind of ground to a halt - there was just so much work to do on the basic infrastructure (mostly interacting with the database, adding properties/fields, and the way rails maps objects to fields, plus synchronization issues, and the lack of a good frontend for the database) that it was kind of discouraging - i need and want something now! maybe rails is the way to go, maybe nitro/og would be better though, i’m not sure at this point. so i don’t know what will happen with the new site design… i know what i want, but i don’t know how to get there yet!

Tags: No Tags

well, it looks like everything is working. although i wound up paring down the export features until it’s pretty bare bones, just because getting the simplest things to work is such a hassle in c++, and takes 10x as long as you’d expect it to. i really don’t want to work on this c++ version any more - i’d rather view it as a prototype for a ruby version.

anyway, all it needs at this point is one more bug fix (the printing extra blank pages one), and adding some things to the help file, and then building the setup program. so it should be done tomorrow. i’ll hold off on the site switchover till a few days after that. and then i’ll probably just put the c++ code on sourceforge - they’re adding subversion service soon -

“The SourceForge.net Service Operations Group has completed the implementation of the Subversion service offering on SourceForge.net. The Subversion service has been made available to a select group of beta candidates for a beta testing phase that began on 2006-01-12. Pending successful completion of the beta phase, the full service offering will be made live in February 2006.”

Tags: No Tags

the next version is nearly done - the main new features are file and folder property links, export data to csv and tab delimited, and using an ini file instead of the registry. plus a few other minor features and bug fixes. it might get out tonight, or this weekend. it’ll be version 1.2a, though it doesn’t include all the things slated for version 1.2 in the old pdf reports.

if i were still using microsoft access i could easily provide a new version of those reports, but alas, i switched the data to mysql, and for whatever reason access does NOT work like it’s supposed to, accessing an odbc data source. it just bombs, badly. it’s apparently just some bug they’re content to leave in there. so the days of easily editing data and generating reports is over, until i either build decent tools with the ruby/rails version or find some good free ones or (ugh, gasp) buy some…

i also need to mess with dns servers and get switched over to the new web server, so the site might go down for a bit (24-48 hrs).

Tags: No Tags

well, the carpet finally arrived, after 4 or 5 schedule changes and miscommunications (not a good recommendation for home depot, unfortunately), and today is when i get all my stuff moved in. and then i’ll be surrounded by boxes until i get it all squared away. i’ll try to at least keep the kitchen clear so i can do some work there. i was hoping to get 1.1e released before this but the usual unexpected complications keep arising. like this one, checking some code in to the repository…

C:\work\neomemLite>svn ci -m “various small features, eg insert date/time”
Sending BDataDate.cpp
Sending BDataHyperlink.h
Sending DialogExport.cpp
Sending DialogExport.h
Sending FrameChild.cpp
Sending NeoDoc.cpp
Sending NeoDoc.h
Sending NeoMem.rc
Sending Resource.h
Sending ViewRtf.cpp
Sending ViewRtf.h
Sending build.rb
Sending hlp\NeoMem.hm
Sending resource.h
Sending version.rc
Sending version.rct
Transmitting file data …………….
svn: Commit failed (details follow):
svn: Checksum mismatch for ‘C:\work\neomemLite.svn\text-base\Resource.h.svn-base’;
expected ‘b6b63cf47af336c772dac8804542bdd6′,
actual: ‘8ec660fc13f4764fc3b80485e23b6cb9′

i have no idea what this means… a corrupted repository? something less ominous, hopefully?

anyway, i haven’t replied to emails or the forum very much lately because i just haven’t had the energy for it - dealing with this move and working on neomem has taken up most of that. and for whatever reason the feedback over the last couple of months has kind of skyrocketed. i appreciate all the comments and suggestions, but i should be better able to respond to them in a week or so, once things start settling down…

update: i wound up just deleting that file from the repository, and readding a backed up copy of it. there’s apparently some way of going in and correcting the checksum error, but i didn’t have the energy to deal with that. i’ve heard that the berkeley db backend on subversion is kind of flaky, and maybe this is an example of it (?). and i’m not even using a server or anything - it’s just me on my laptop.

Tags: No Tags

It’s the move that never ends - I’d planned to stay at my cousin’s for a few months, but I’m finding it really hard to get work done here, so I found an apartment, and will be moving in there this week. In the meantime, the code for export is moving along, slowly…

Tags: No Tags