BYTE Magazine

[Keatah]

I respectfully say that some of the ones in the torrent have washed-out covers. Brightness + compression way too high. And this also shows up as a fringe around some advertisements elsewhere in the mag. OTH, the marcus scans are too dark, but some of the covers are better. And some of these scans even have OCR text replacing the scanned text - too harsh on the eyes.

 

As the repository grows I'm grabbing it all and picking and renaming them in a nice & sortable way. Straight from Windows Explorer. At least something that makes sense to me. Like TOSEC style. As it stands right now, there are 3 or 4 naming schemes on the ftp site. And we don't know what scans are from who.

Edited March 14, 2013 by Keatah

[mariush]

It's interesting how you're so vocal about something that's free and takes a lot of time to accomplish. Someone's wasting hours of his free time to scan, perform ocr and upload an issue of a magazine just so that you would have something to complain about.

 

I suppose it would be better for him to stop completely.

 

If you're so picky, why don't you go buy a 50$ scanner and a few magazines from eBay at 20-50$ a piece and post them here? What, you don't like things unless they're free?

[Keatah]

Something like that. When it comes to the technology industry and computing; it owes me big-time.

[Keatah]

So like when are more magazines going to be properly scanned?

[Keatah]

First off. We have the ThumpNugget scans - these are overall good quality and generally my first choice for reading. While they are ever so slightly overexposed, there's little to complain about considering the wide dynamic range of text and artwork. The quality is pretty even and consistent across the collection. An excellent job here! Too bad they put a stop to it. ftp://helpedia.com/pub/archive/temp/Byte/

 

Next up is the batch of Marcus scans. These are all over the place, errors abound. Missing pages. Double-fed and skewed pages. Under-exposed graphics. See-through paper. Marked up covers. Some kinda gay-ass pdf security if you want to edit them or replace pages (but you can work around that). The two redeeming qualities are that the text fonts are easy to read, similar to the ThumpNugget scans, and the fact that the scans are pretty much un-retouched or overcompressed by the scanning software.. These also retain the look and feel of the magazine. Like the [t] scans, these [m] scans show the style of the font and graphics. Too bad about the under exposure and carelessness in the actual scanning. And also, be aware of the near GB size for some issues. ftp://helpedia.com/pub/archive/temp/Byte/Other%202/

 

And then there are the "other" scans. These are difficult to read, you'll not find them pleasant to look at for long. And there's a lot of blown out areas. You'll not see the fine details in Byte's beautiful covers. Nor will you relive the 70's typeface or advertising images in all their glory. These [o] scans are small in size and suffer from too much automated retouching. The graphics have a smeared wet watercolor look when compared against the other two. There seems to be too much sharpening and contrast enhancement on much of the text. In my opinion these are just as bad as the Marcus scans. Just in different ways. And it would be a good thing to replace these as well. The redeeming quality here is they seem to be more complete with less missing pages and things like that. These would benefit well if the compression and "enhancement" weren't so aggressive! ftp://helpedia.com/pub/archive/temp/Byte/Uploads%202012/

 

Then there are the misc scans in the other directories that I didn't really analyze yet. It is important to learn how to use your scanner correctly. Not just feed paper in and blindly post the results. Despite what advertising would have you believe, scanners and their software are still pretty primitive. I only criticize harshly in hopes that improvements will be made.

 

But, anyways, I propose this is how a collection should be labeled and organized.

[Jim02762]

Something like that. When it comes to the technology industry and computing; it owes me big-time.

Just remember that nobody here owes you squat.

 

So like when are more magazines going to be properly scanned?

For you? As soon as you get to it.

 

For the rest of us there are a couple new ones just uploaded.

 

First off. We have the ThumpNugget scans - these are overall good quality and generally my first choice for reading. While they are ever so slightly overexposed, there's little to complain about considering the wide dynamic range of text and artwork. The quality is pretty even and consistent across the collection. An excellent job here! Too bad they put a stop to it. ftp://helpedia.com/pub/archive/temp/Byte/

He certainly did a very good job. I guess he just didn't have the time to continue. But it would be nice if he'd stop by and say hi.

 

Next up is the batch of Marcus scans. These are all over the place, errors abound. Missing pages. Double-fed and skewed pages. Under-exposed graphics. See-through paper. Marked up covers. Some kinda gay-ass pdf security if you want to edit them or replace pages (but you can work around that). The two redeeming qualities are that the text fonts are easy to read, similar to the ThumpNugget scans, and the fact that the scans are pretty much un-retouched or overcompressed by the scanning software.. These also retain the look and feel of the magazine. Like the [t] scans, these [m] scans show the style of the font and graphics. Too bad about the under exposure and carelessness in the actual scanning. And also, be aware of the near GB size for some issues. ftp://helpedia.com/pub/archive/temp/Byte/Other%202/

There are more problems than just those. And I'm looking at Marcus' Oct. 1985 and the last thing it is is easy to read. It's so washed out I can barely make out the text.

 

The PDF security is very easy to get around; it's a null password. Nothing, at least on the one(s) I looked at awhile ago.

 

And then there are the "other" scans. These are difficult to read, you'll not find them pleasant to look at for long. And there's a lot of blown out areas. You'll not see the fine details in Byte's beautiful covers. Nor will you relive the 70's typeface or advertising images in all their glory. These [o] scans are small in size and suffer from too much automated retouching. The graphics have a smeared wet watercolor look when compared against the other two. There seems to be too much sharpening and contrast enhancement on much of the text. In my opinion these are just as bad as the Marcus scans. Just in different ways. And it would be a good thing to replace these as well. The redeeming quality here is they seem to be more complete with less missing pages and things like that. These would benefit well if the compression and "enhancement" weren't so aggressive! ftp://helpedia.com/pub/archive/temp/Byte/Uploads%202012/

I have no idea why you're having so much difficulty reading these and I'm not. I agree that his settings need work and in the past couple of days he's done wonders with it. But I have no idea how you could even begin to compare these with the Marcus scans. I can (apparently magically) read these with no problem, actually see what's in the ads, and even have ALL the pages, in order, in the articles I'm reading. There's nothing I can't see. It's just sometimes it could be sharper.

 

If he's held on to the original TIFFs and could re-process them using the settings he's been using the last couple of days then that would be awesome. But that's a lot of TIFFs to be holding on to...

 

Then there are the misc scans in the other directories that I didn't really analyze yet. It is important to learn how to use your scanner correctly. Not just feed paper in and blindly post the results. Despite what advertising would have you believe, scanners and their software are still pretty primitive. I only criticize harshly in hopes that improvements will be made.

Oh is that why you criticize so harshly? You're tough love leaves a lot to be desired, kid. Show some respect to people that are actually investing considerable effort in a rather difficult task. I think it's quite obvious this person isn't blindly loading pages into a hopper and hitting 'Scan'.

 

But, anyways, I propose this is how a collection should be labeled and organized.

Not everyone uses Windows. That would be a nightmare to use from a UNIX command line. For example, BYTE\ 1977\ \[01\]\ January\ -\ Hash\ Tables\ and\ Interrupts\ \[t\].pdf . Whereas 197701\ BYTE\ 02-01\ Hash\ Tables\ and\ Interrupts.pdf is slightly better but is easy to use with tab-completion. (underscores were almost used instead of spaces)

 

And what's in the first set of brackets??? Those aren't issue numbers? And what about volume numbers?

 

But if you want to do it like that then you go right ahead. Nobody is stopping you.

 

 

To the guy/gal currently doing the scanning, *I* think you're doing great. I see you're working on the colors and they look, to me, a million times better. Going without ClearScan has made them a lot larger but, personally, I prefer it for a couple different reasons. But I did notice the June 1988 issue you just posted had a problem with the Apple ad on PDF page 37. Looks like some of the bottoms got clipped. Other than that it's looking pretty good to me!!!

[akator]

blah blah blah

blah blah blah

 

If the purpose is to scan material for archival purposes, the highest quality reproductions possible should be made (within reason).

 

If a job is worth doing, it is worth doing right.

[Jim02762]

If the purpose is to scan material for archival purposes, the highest quality reproductions possible should be made (within reason).

 

If a job is worth doing, it is worth doing right.

I completely agree.

[Hatta]

But still, a washed out scan is better than no scan at all. And they can always be replaced later. I appreciate everyone who's contributed to this project.

 

That would be a nightmare to use from a UNIX command line. For example, BYTE\ 1977\ \[01\]\ January\ -\ Hash\ Tables\ and\ Interrupts\ \[t\].pdf . Whereas 197701\ BYTE\ 02-01\ Hash\ Tables\ and\ Interrupts.pdf is slightly better but is easy to use with tab-completion.

 

Do people still use shells without tab completion? As someone who manages files exclusively from the UNIX command line, I don't mind the square brackets so much, they can be handy for globbing/regex, and they're no big deal if you type your filename in quotes instead of escaping everything. I do however prefer ISO format for dates, which would be 1977-01. BYTE should come first, as it's the most general descriptor. And yeah I think the volume/issue no. should be preserved in the filename.

 

So my ideal filename would be something like this. "BYTE 1977-01 [02-01] - Hash Tables and Interrupts [t].pdf"

[Jim02762]

But still, a washed out scan is better than no scan at all. And they can always be replaced later. I appreciate everyone who's contributed to this project.

I completely agree with that too.

 

In my opinion, the information contained within the BYTEs is the most important thing. Its presentation/appearance is a close second. I'd gladly take a crummy but legible photocopy if that's all that's available even though what I really want is a high quality scan.

 

Do people still use shells without tab completion? As someone who manages files exclusively from the UNIX command line, I don't mind the square brackets so much, they can be handy for globbing/regex, and they're no big deal if you type your filename in quotes instead of escaping everything.

I work from the command line almost exclusively. Tab completion is a must nowadays. Anyone using bash or zsh has it. Not sure about the older ones like csh and ksh. I just wish the Windows shell (or whatever it's called) had it. It probably needs tab completion more than any other shell!

 

Square brackets in a name I find annoying because I do use them for globbing. I think in a UNIX environment they should be used in file names sparingly to avoid confusion.

 

The only problem I have with using quotes is I always forget to type in the opening quote until after I type/paste the file name. Every single time. lol You also can't use globbing within them obviously.

 

I do however prefer ISO format for dates, which would be 1977-01. BYTE should come first, as it's the most general descriptor. And yeah I think the volume/issue no. should be preserved in the filename.

 

So my ideal filename would be something like this. "BYTE 1977-01 [02-01] - Hash Tables and Interrupts [t].pdf"

 

The reason the dates are first is so that a YYYYMM<tab> will almost always be all you have to type to get the complete file name to pop up. This is very handy when trying to sort out what's what and PDFs are getting moved all over the place. However, I suppose for most people just downloading and reading them that's much less of a concern.

 

YYYY-MM probably would not be a bad idea. Just as easy to type and easier to read. Torrents probably should have been done like that.

[akator]

In my opinion, the information contained within the BYTEs is the most important thing. Its presentation/appearance is a close second.

 

Because it is 2013 and we're talking about 20+ year old content, I think the presentation (including images, appearance, design, ads, typefaces, all of it in entirety) is just as important as the text. The "raw information" is no longer necessary... we are reading and enjoying these because of the history, not the "raw information."

 

More importantly, because it is 2013, and digital cameras over 10MP are commonplace, there is no reason for bad imaging of any kind. Books and magazines don't need to be destroyed anymore to run through flatbed scanners. Photo editing tools to rotate pages for correct alignment are open source and free.

 

I'd gladly take a crummy but legible photocopy if that's all that's available even though what I really want is a high quality scan.

 

If we were living in 1993, I would agree with you

[Jim02762]

Because it is 2013 and we're talking about 20+ year old content, I think the presentation (including images, appearance, design, ads, typefaces, all of it in entirety) is just as important as the text.

When I said that presentation was a close second, I meant a really close second.

 

The "raw information" is no longer necessary... we are reading and enjoying these because of the history, not the "raw information."

That's your opinion and you're certainly entitled to it, but I disagree. Just because BYTEs are old doesn't mean the raw information is no longer necessary. It depends on who you are, what you're doing, and what you consider "necessary".

 

I read BYTE now to enjoy the history AND for the raw information. Maybe I'm in the minority.

 

More importantly, because it is 2013, and digital cameras over 10MP are commonplace, there is no reason for bad imaging of any kind. Books and magazines don't need to be destroyed anymore to run through flatbed scanners. Photo editing tools to rotate pages for correct alignment are open source and free.

Wish I had one. And I wasn't aware scanning had become so easy to get right. Good to know.

 

I do have to get a decent digital camera and build a jig. I have some old magazines and books to scan and can't bring myself to cut off the bindings.

[Keatah]

Well, in any case I enjoyed the May-1992 issue. Reading about the clock doubled Intel 486DX2-50. In fact I was so enthralled with it back then I went out and got one. Still have that machine today.

[Jim02762]

Well, in any case I enjoyed the May-1992 issue. Reading about the clock doubled Intel 486DX2-50. In fact I was so enthralled with it back then I went out and got one. Still have that machine today.

That was my first 486 chip. Excellent CPU. I put it on a 386/486 hybrid board and used it for many, many years. Got a 486DX4-100MHz to replace it and tried running it in DX2 mode but it wasn't reliable for some reason.

[mariush]

A forum member has contacted me and gave me an issue of Byte to publish on the FTP server. As he didn't mention if he wants to be anonymous or not, I'll leave it at that.

 

The file is located in the /March 2013/ folder. Had a brain fart and couldn't think of any better name so live with it, I guess.

 

He says he has two other issues of Byte that he's willing to scan and wants feedback regarding the current issue he produced for us.

 

He still has the raw files and Abby Finereader 11 for this issue and he seems motivated to redo the issue taking your feedback in consideration.

 

ps. Download speeds may be slow for a while , as the server is busy with other stuff so don't panic if the download speed is low, it will get better,

Edited March 23, 2013 by mariush

[high voltage]

Where's the 92 issues, please?

[Keatah]

I don't see any /march2013/ folder.

[mariush]

If you're in "Uploads 2013", go back one folder and you'll see that folder.

 

ftp://helpedia.com/pub/archive/temp/Byte/March%202013/

 

[Keatah]

Bwahahahaa!!! Ridiculous!!! I failed to refresh my browser cache..

[Jim02762]

He says he has two other issues of Byte that he's willing to scan and wants feedback regarding the current issue he produced for us.

 

He still has the raw files and Abby Finereader 11 for this issue and he seems motivated to redo the issue taking your feedback in consideration.

I'm having slight issues with this PDF in Linux with both Adobe Reader and evince. Doesn't seem to always render correctly. Not sure what's up with that.

 

I think the scanning may be pretty good but the OCRing is killing it for me. Finereader is trying to replace all the text with vector graphics, just like Adobe's ClearScan (which I prefer off) but not as good. There may be other post-processing problems too. For example, on PDF page 13 the resolution of the guy seems fine but the title at the top seems a little blurry, the fonts don't look that great, and the box and floppy disk to the guy's right seem blurry and have artifacts. But I can't tell if this is because of the problems with my PDF readers, post-processing/OCRing, or what.

 

My feedback would be to get ABBY Finereader to not touch the original images at all. Don't let it replace the text because it's not very good at matching the fonts.

 

Also, don't let it automatically create a PDF table of contents. It's all messed up.

[Jim02762]

Where's the 92 issues, please?

The one Keatah mentioned with the 486DX2? Should be in ftp://helpedia.com/pub/archive/temp/Byte/Uploads%202012/ .

[Keatah]

Some feedback for march 2013 - It has a great balance in how the software settings are configured.

A forum member has contacted me and gave me an issue of Byte to publish on the FTP server. As he didn't mention if he wants to be anonymous or not, I'll leave it at that.

 

The file is located in the /March 2013/ folder. Had a brain fart and couldn't think of any better name so live with it, I guess.

 

He says he has two other issues of Byte that he's willing to scan and wants feedback regarding the current issue he produced for us.

 

He still has the raw files and Abby Finereader 11 for this issue and he seems motivated to redo the issue taking your feedback in consideration.

 

 

 

Well, I didn't look through every page for completeness or consistency. But while spot checking several pages I noticed some anomalies.. Here are the PDF page numbers:

 

Cover: Excellent! Source material imperfections and marks abound! I wouldn't have it any other way. Little or no compression artifacting. Great framing and cropping. In fact, VERY NICE gamma and dynamic range on the color pictures.

 

Many pages: Text is still processed too much, almost as bad as the [other] scans. Too much sharpening and too much contrast around the letters. Too harsh on the eyes. The original Thumpnugget scans have multiple shades of grey for font smoothing. At least 8. Here we have 2.

 

191 is skewed

209 shows mixed compression / no-compression

216 the top 3 lines of text show the compression that is so annoying in much of the [other] scans. But if this is limited to large black text then it's alright I suppose.

4 some unevenness in the density of the first 3 lines of text.

6 too much compression

15 some compression wiping out the details in the graph

18 McGrawHill symbol overcompressed

25 some of the top text is compressed/smoothed, some isn't.

26 way too much compression on the violet area of the printer

27 good readability of the super small text at the bottom. Maybe less sharpening or processing to make it even better?

224 compression not consistent - wiping out advertisement text details is the S&W Computers ad.

226-227 inconsistent & too much compression. Look at the hieroglyphs/icons on the printer control panel. On 226 you can't see what they are, they are compressed into oblivion. On 227 the text is almost readable, and their associated symbols are quite visible!

 

I'd like to see one of the un-processed .tiff images that has a lot of text on it. Perhaps I can educate the guy on how to get the settings correct. Software is too complex for most people and has too many options to play with.

 

At this time the [t] scans remain the gold standard for Byte rags. Though this last one is better than the [m] and [o] scans. Especially with the color images! I feel if the text smoothing is corrected and compression is less aggressive and spotty we'll be in business.

 

You might think me an ass for being critical. You know what? I don't really care. I'm telling you how it is. In the meantime keep'em coming! I want more to read.

Edited March 24, 2013 by Keatah

[Keatah]

Agreed on the oversharpening or vectorization of the text. While readable, of course, it isn't pleasant. Most people are only aware of a slight (if that) discomfort when viewing fonts like this. But truth of the matter is they suck.

[Keatah]

Here.. Which do you like better?

 

o.bmpt.bmp

[DanBoris]

Because it is 2013 and we're talking about 20+ year old content, I think the presentation (including images, appearance, design, ads, typefaces, all of it in entirety) is just as important as the text. The "raw information" is no longer necessary... we are reading and enjoying these because of the history, not the "raw information."

 

I have to disagree with the "raw information" not being necessary. I recently acquired a IMSAI 8080 system with a floppy drive that I could find no information about on the Internet. It also had a floppy disk controller that again isn't document anywhere I could find on the net and used an IC chip that I could find no information on either. Digging though some old Byte issues I ran across an article about how to interface this exact floppy driver to a system using the same unidentified chip that is on the controller I have. This information will be invaluable in getting this system running again.