PlanetDjVu Forum Thread - The Future of DjVu
Published in compiled form on October 20, 2003
On August 1, 2003, we put out an open call for comments on the future of DjVu.
Many of you contributed to the PlanetDjVu Forum Thread. Thanks for the kind words about PlanetDjVu - they are a big reason that this website is still active.
Now, here are your contributions and insights in compiled format:
Should we continue to operate the PlanetDjVu website? Does this site have value for you? Is the DjVu format worth reviving? Please post your comments for review by us all and by the new management of LizardTech in particular. Now is the time to speak your mind!
PlanetDjVu was the first site that I found when I searched for DjVu. It is the only web resource that publishes reliable information (bad or good) about the commercial DjVu technologies.
I learned a lot from this site and its publisher James Rile. I'm yet sure that DjVu is the best framework for distributing document images over the Internet. Its supremacy on PDF is still big (please read about PDF usability on Jacob Nielsen's useit.com - http://www.useit.com/alertbox/20030714.html and the other emerging image standards, like JPEG 2000, are new on the market and lack efficient viewing and distributing technologies.
The new Lizardtech Owner could be a big opportunity to open a new era
for DjVu. Please Jim, don't give up!
Wish list for DjVu new owner:
--affordable developer license;
--new business model for final user (page cartridges model is not working);
--beta testing and support for free viewing software;
--considering the use for background layer of JPEG 2000 standard;
--support the open source companion project (very important and strategic);
--developing a strong copyright system protection
Too bad you stop but I understand it must be really discouraging. Let us see what Celartem is going to do with DjVu... Anyway, thank you for all the valuable information you provided about DjVu and other subjects.
As an ex-lizard who did not exercise his options and feels no remorse about it, I find planetdjvu an interesting place to visit from time to time.
My main interest in the site is from Jim's informative posts about Lizardtech's incompetent handling of DjVu, from technical glitches to rants about the moronic management at Lizardtech.
Jim seems genuinely interested in the format and has many good ideas about keeping the format alive. To the new management, I would say that having this man on your side would be of great help in your attempts to get DjVu on track.
He has also given a lot of useful technical advice to people who were struggling with Lizardtech's products.
1. by default built-in viewer in Internet Explorer
2. java viewer
3. on-screen OCR editor
4. more layers ( by using additional external plug-ins )
I for one very greatly hope that PlanetDjVu continues to exist. Your forums are an amazing resource for the users of DjVu.
I also think the DjVu format is here to stay no matter what. It is simply too good to die. The "indirect" format of DjVu makes it so far ahead of everybody else in terms of delivering web content efficiently that it has to be only a matter of time before it takes off like a rocket. After all its inevitable that thousands (10's of thousands) of libraries are going to put collections on-line in the next few years. And that's a market worth 100's of millions waiting to be picked up by the owners of DjVu!
Almost every library site you visit on-line says they have a funded digital library project going, but almost none of them have anything completed to show. Yet you can take a scanner hook it up to a pc and scan a 300 page book by hand in a couple hours, run it thru DjVu Solo in a minute or two and you've got an on-line book.
I first discovered DjVu just over a year ago, not through LizardTech's largely invisible marketing, but through James Riles' articles and later through PlanetDjVu.
DjVu impressed me as an exciting and innovative technology, but for me the technology was less interesting than the range of compelling opportunities which came to mind with a high performance, compact, full colour imaging standard. Finally a format which did not assume the world was on broadband. Finally something which made full colour imaging practicable.
I was (and still am) under no illusions about what it takes to make build a new standard - especially in light of the almost overwhelming dominance of PDF. I nevertheless didn't for a minute expect that LizardTech would be the first and biggest obstacle.
The arrogance and incompetence displayed by LizardTech should be studied at business schools the world over as a case study on how not to commercialize an innovation. My dealings with them have been beyond frustrating - at best I could describe them as unresponsive and unsupportive. It appeared to me that they were so fond of DjVu they decided not to share it with anyone.
A little excitement goes a long way, but it was patently clear that DjVu had no real chance of success with LizardTech at the helm, and I decided to focus on more realistic things (while still hoping that something positive might one day emerge.)
This new development may well be that positive step, but any future for DjVu is very different one from one year ago. If something is to be salvaged and built on it needs to be immediate, decisive and even aggressive - more of the same will further relegate DjVu to the footnotes as a once promising technology which never quite made it.
Finding a way to embrace the positive energy of people like Jim, and others like myself who have expressed real interest would be a good start. How about appointing Jim to an advisory board and reconciling with the original developers?
How about simplifying the licensing and making it really attractive? How about encouraging supporting products while remaining custodian of the quality (which has been left to stagnate within LizardTech.)
Jim - don't give up just yet;) PlanetDjVu is a great resource and one which deserves to succeed along with what is still a technology full of potential (but that potential needs people to unlock it, and LizardTech needs to demonstrate that it understands this.)
YES. As a user of DjVu, I have now converted 1.5 million pages into DjVu format and am very keen to see this technology succeed. We need this format to be far better marketed in the UK and Europe so that it reaches its full potential.
PlanetDjVu can only help to achieve this and if in financial difficulties should receive some support from LizardTech! Keep up the good work.
I've been working with DjVu since 1999, and it's my expectation and hope that the format will flourish in the future. Having worked with the original developers at AT&T, I can vouch for DjVu's splendid "pedigree", and working now at the Rutgers Library I can vouch for the intense interest in DjVu in the academic digital library world - especially perhaps in the open source DjVuLibre development. They love the format.
Perhaps the one thing that has been holding some people back from taking the plunge has been worry about whether Lizard Tech was about to go "belly up" as one woman here put it. Everybody uses PDF, but nobody loves it they way people love DjVu, and my feeling is that with a little better support that LT was giving the format, many digital library projects would adopt it.
My wish list would include better support for Macintosh (always a sort of step-child for DjVu developers, but still important among academics and graphics people). It would also be nice to see the DjVu (the Java viewer) project at SourceForge take off. PlanetDjVu has an important role to play, especially in support of the Windows/Mac side.
Yes please. We use DjVu for many of our projects. To list a few: The Spectator, Latin and Greek Grammars and Dictionaries, The Electronic Journal of Boundary Elements. We use DjVuLibre quite often and would like to see it continue and if possible as open source. We like to look at the code. DjVu is superior to any other presentation format for the work we do and would certainly like to see it flourish.
I was heartened to learn about the "new life" possible for the djvu document type. I spent months researching methods for converting our newspaper clippings collection for our state agency and discovered the DjVu format. Unfortunately, the format was constrained by its licensing that dampened open source development, 3rd party tool development and even distribution of a reader using SMS-like network upgrades.
I was drawn to the format for several technical reasons, one being its highly-compressible nature. I am confronting a huge digitization project and document size factors extrapolated over a massive collection are a significant consideration. We had other requirements for the project, including adding watermarks and file-splitting that we found to be well-supported with the DjVu format.
Digitization may seem to be well-along -but actually there is still a huge need for innovative document imaging/ storage and retrieval methods. I would like to see the DjVu format flourish by making it reasonable available to developers and end-users.
Bill C. Riemers
It's dead Jim.
I really would like to see DjVu revived, but I just don't believe it will ever happen. The golden window of opportunity was 3 years ago, when most internet users still connected via dialup. Adobe keeps upgrading and improving their product. DjVu has been stagnant for too long to even play catch-up.
There just is not the interest out their even for free products. For example, DjVuLibre needed just three votes to be added into the Cygwin distribution. Anybody in the world could vote. After about six months, it received only two.
Evan as one of the original developers, I no longer use DjVu. Certainly, I still have DjVu files on my website, but I haven't been adding anymore, and I offer all the DjVu documents in alternate formats.
Still, I do wish for you to continue to operate this site, as I still find the updates interesting, and I keep hopping something will change for the better.
Yes, I agree that the golden window of opportunity has passed, and that precious three-year window of opportunity was largely wasted and squandered by LizardTech. But DjVu is not necessarily dead. There is hope that Celartem will completely reform and restructure the way that DjVu is presented and licensed.
There is a new owner, and a new CEO, with no ties to the decisions of the past made by the venture capitalists that previously ran LizardTech.
If LizardTech and DjVuLibre work together now (with ACTIVE support from LizardTech), the results could bring about the revival that is needed. For example, volunteers at DjVuLibre can make the Java viewer, while LizardTech makes a better Windows editor. It will become easy (and affordable) for DjVuLibre users to license the segmenter from LizardTech.
I think that your Cygwin distribution will do better once DjVu (and confidence in LizardTech) is better.
The key to the revival of DjVu, I think, if for Carlos to take a leadership role in this. For one reason, this is the only way to quickly restore confidence in LizardTech. For another, LizardTech owns the format and controls the commercial licensing, so even the best efforts of open-source people and third-party people will not turn the tide, without the active support of LizardTech.
My intent now is to give Carlos the opportunity he has asked for, the "fresh-start" for DjVu, and we will see if it happens.
>I do wish for you to continue to operate this site, as I still find the updates interesting
Let me conclude with a remark about updates. I have posted at least one update per week over the past two years. DjVuZone.org has had no updates, and those at LizardTech have been sparse. There were no published articles on DjVu last year outside of PlanetDjVu. It could well be that DjVu has been suffocated by a lack of information, and that simply supplying information will contribute greatly to its revival.
DjVuZone.org in particular turned from an amazing resource for DjVu into a cancer. Because it is now so out of date, it is proof-positive that DjVu is a troubled format.
When I introduce people to DjVu, they often say "I never heard of it". I then show them PlanetDjVu and they say "is it your format?". I hope that both of these questions will go away, as a result of others beside myself providing informative updates on what is happening with DjVu!
Finally, a little note to Carlos, for when he reads this: Bill Reimers is a brilliant technical resource that you may well wish to call upon!
"It's life, but not as we know it..."
I don't agree with you Bill. The Internet connection speed in the rest of the World (I'm living in Italy) is grown but there are a lot of old modems still installed and working.
My opinion is that band-saving is yet an important issue because the resources are always limited. That's the reason why there is a strong research activity for developing compression algorithm for video (mpeg), music (mp3), and still image (jpeg2000, JBIG2).
Why I can download a 25 MB TIFF page file, even with an 2mbit speed optical fiber, when I can get the same quality with a 200kb DjVu (this is real ratio for a 400dpi 256 greys A3 TIFF File)? On the other side there are viewing problems. I tried to display a 25 Mb tiff file even on two processors PC with video card accelerator, I'm not able to get the instant rendering effect of DjVu.
In my opinion the real challenge for DjVu is the emerging of other efficient compression and file format standards like JPEG 2000 and JBIG2. When a tech company will able to join the two technologies (that are standard), and develop an efficient delivering system, then DjVu will have to work hard to find its place on market (also because its technologies are not standard).
I've been watching DjVu for years as an even more efficient alternative to the Xerox .xif TIFF-FX formats, but have been hesitant to make the switch due to DjVu's apparent instability. Now, that ScanSoft has dropped all support for the .xif format...I am very, very interested in finding an efficient and long lived standard for my gigabytes of archives.
Planet DjVu is my *only* source for competitive DjVu information, and it will be of vital importance in reaching a decision for my next standard.
The DjVu standard is not presently in use here. That could change rapidly, if I see some real efforts at consensus/standard building. This whole XIF episode has left a very bad taste....and I'm seeking ways to avoid reoccurrence.
For all intents and purposes PlanetDjVu is DjVu. Nothing is more impressive than that Australian newspaper site with the front page in Digitial DjVu. I have never seen anything as impressive in PDF.
Since the future is in small devices with low-power CPUs, etc. it seems that DjVu would have a huge advantage over bloatware like Acrobat/PDF.
But lots of rough edges remain. Harnessing the Open Source community and leveraging what they are already doing and porting that to Windows seems to be the way to go.
JRAPublish beats the competition and could make a great companion tool to convert DjVu to PDF and back again so we can have the best of both worlds as our needs dictate.
Arnold is talking about wentworthcourier.com.au" www.wentworthcourier.com.au, where Prepress PDF newspapers are converted to DjVuDigital and look stunning in the hosting environment created by Realview Technologies.
This is the only site on the web so far with a complete DjVuDigital collection, so the planet has generally not seen DjVuDigital to appreciate it. As demostrated, it can effectively re-purpose prepress PDF to make small, web-deliverable DjVuDigital files.
The cost to get licensed to produce DjVuDigital files with the latest software has been about $10,000 in the past from LizardTech, and this is one reason that DjVuDigital is not appreciated. The folks at DjVuLibre want to see the DjVuDigital code to go into open source, to be included with DjVuLibre, in order to "kick-start" DjVu back to health again. These are two extreme positions for DjVuDigital, and they need to be reconciled by the new management of LizardTech.
Again, the enormous potential of DjVuDigital is that you can create a graphically-rich publication using your favorite authoring problem, such as Quark, and you can then produce a PDF file for the print job, and you can then convert that PDF file to a compact DjVu file for web delivery. One job, two forms of published output: physical and digital.
JRAPublish can convert from PDF to DjVu by first rasterizing the PDF pages, and then using image segmentation to make DjVu pagess, but this is not the same as DjVuDigital conversion, which uses all the objects within the PDF to perform a superior form of "DjVuDigital" segmentation. We would like to someday offer DjVuDigital conversion in JRAPublish, so that complete existing archives of PDF files (of all flavors) can be batch-converted to DjVu.
A reminder that a beta version of DjVuDigital conversion software is used in the
Any2DjVu conversion server. Just submit a PDF file there to see the amazing potential of DjVuDigital conversion for yourself.
Economics and convenience ensure a future for online viewing of print formatted documents-- both color and b&w. Without question, DjVu is technically superior to PDF for color. PDF dominates all online docs, however, because of Adobe's superior marketing strategies and deep pockets to keep it rolling.
If DjVu is to find a significant niche for itself, I believe that it needs more than better compression. People use PDF, but Acrobat's clunkiness often makes it painful. DjVu's Indirect architecture is an improvement to Acrobat, but I doubt it is the legs that DjVu needs to catch a real market.
One way or another, DjVu needs to mesh better with HTML to survive. Whether this can be accomplished through DjVa, licensing with MSFT, Google -- I don't know. But I do believe that DjVu's faint pulse could be revived if it lent itself better to *browsing*.
Attention spans on the web are short. Acrobat is a clumsy pig suited for pre-press and web distribution of IRS forms, not casual browsing of a glossy magazine via a website or email thumbnail. There is an opportunity for leaner and cleaner color doc display -- maybe Celartem will take advantage with DjVu, or LuraWave with LuraDoc.
Maybe the answer isn't a single proprietary format, but rather XML representations such as Texterity's sans plug-in Textcafe. Or maybe Adobe will call off all bets by introducing different flavors of PDF and Acrobat for different apps.
This site is a tremendous resource for DjVu and what's left of its community. The end of the site would be a big blow to Celartem, whatever it is that they have in mind for DjVu.
I suggest that the site be maintained long enough for Celartem to show us what they can or can't do-maybe another year. Dynamic people with proactive and reactive capacities could change the web and make a whole bunch of money with DjVu.
Many thanks for the valuable service that you provide with the various aspects of the Planet DjVu site. We believe DjVu to be an excellent way of making 120 years of archaeological papers relating to Gloucestershire, UK, available to all on the Internet.
However, we have been very concerned, like many others about the policy of 'per page' charges rather than a flat one-off payment for a licence. We wait to see what the new owbers of DjVu will come up with. We hope you might be able to keep the PlanetDjVu site open (perhaps in a more limited form) until the situation becomes clearer.
Ray Wilson, Gloucestershire, UK
Rodrigo Barbosa e Silva
Hello friends from PlanetDjVu.
It's a great chance to opinionate about the future of PlanetDjVu and DjVu. I guess that DjVu is a strong format to deal with digital documents. Brazil is a big market to new technologies, the majority of enterprises are looking for a solid and powerful solution. To speak about the future of DjVu in global marketplace, I have to explain about brazilian perspective.
The main problem to adopt DjVu in Brazil is the price. We have an indexed economy, it means that dollar every day has a different value. For example, today US$ 1.00 is bought by R$ 3,03. Last week, US$ 1.00 was only R$ 2,90. I do not know how will be the currency next week, next month and so on... I guess the better form to expand DjVu in Brazil is adapting the products and all interfaces to Brazilian Portuguese.
Another point is to establish the cost of products in Reals, so when dollar goes up, the DjVu won't goes up automatically together. The same when dollar goes down, DjVu do not need to go down.
About PlanetDjVu itself, it's the most complete space to follow the DjVu news, updates and ideas. PlanetDju.com is really important to DjVu Community and I hope that it will have a long life.
I believe in DjVu and PlanetDjVu and I hope that in a short period we will make business in Brazil with this revolutionary technology.
I forgot to mention the one feature that makes me choose DjVu over PDF everytime: custom printing sizes.
In DjVu I can view the document at any percentage of the original size say 125%, and then choose "Current Zoom" as the print size and get a page that is 125% of the original size on paper.
This is crucial to some applications and yet there is no way to do this in Acrobat or any PDF viewer I have seen.
Thank you DjVu for including this simple yet very important feature.
I feel the only way the DjVu format will ever become commercially successful is to completely overhaul the current product strategy by Lizardtech. Over the past couple years, DjVu has been positioned as an "enterprise" solution, yet the stark reality is that the technology as it exists today is better described as a utility that complements larger custom built or packaged document systems. This is in no way an enterprise product in and of itself, and therefore can't be sold and priced that way.
The allure of pricing DjVu at enterprise price points contributed to the out-of-touch pricing / delivery model that has been hugely unsuccessful. I suspect a good deal was positioning to sell LT with less regard to actually getting the marketplace to adopt DjVu. Lizardtech's business model should be more like Lead Technologies (www.leadtools.com), a successful imaging tools vendor. To succeed LT must consider the following actions:
1. Convert the standard encoders to a online low cost purchase model. The basic desktop encoder w/o royalty items such as OCR & annotation, might even be very low cost or even free with an option to purchase the "plus" encoder for maybe $50 and get OCR, annotation, etc.
Must be able to download and purchase in realtime, plenty of sites can outsource this for quick time to market. Look at .com" www.995pdf.com and view their customer list; ie. Ford Motors. They are no doubt generating more revenue w/ a inferior $10 PDF encoder than LTI is with $10k DjVu encoders! The average home user, home business, small business is not going to pay $200+ for Acrobat, this provides an alternative with even higher quality. If LT has no resources to pull this off, I bet the PlanetDjVu can figure this out in short order.
2. Did anyone see the PC Mag test recently of 3rd party PDF encoders? What if LT or another 3rd party had provided a DjVu product, that was reasonably priced? My guess is that it would have gained a side bar due to the quality & performance.
Note: The BlackIce firewall products were offered free for home users initially, and when John Dvorak in PC Mag recommended them in his column (several yrs back), the demand / download traffic was huge and a successful product was launched. Send every magazine writer a free copy!
3. Make it affordable for non-profit and education where the some of the strongest supporters exist. I'm talking truly affordable here.
4. Banish the pay as you go model. Too much development time has been wasted on it, customers don't like it, and when you're trying to build marketshare that's absolutely insane. The folks at Lizardtech hang on to this notion due to the pie in the sky projections of millions of pages/data being encoded but it's a shell game and should be abandoned.
5. Support the open source community once again and start actively and affordably licensing the technology to those who are willing to do something with it. LT has few resources left, and DjVu being kept exclusive guarantees it's death and means Celartem bought MrSID and GF essentially.
P.S. I am a former LizardTech employee laid off in Feb of this year, now on to other things, but who would still like to see the DjVu format succeed & believe it can given major shifts in strategy.
Yes. *None* of the current document formats really work - so DjVu is still very much in the running.
Also, while I know that the Xerox .xiff format is a subject of some derision here, Scansoft's large user base is now very much up for grabs. All it would take to capture this significant user base would be some efficient OCR & archival conversion tools.
I have gigabytes of .XIFF files and nowhere to go .....
I hope that you will continue with Planet-DjVu Jim. I've been using this technology since I saw it introduced by AT&T at an AIIM show in Atlanta and I've not found any other option that is as inherently flexible as DjVu. The major problems with this format have been related to finance and security issues--it was launched into a faltering economy to a user-base that has become very wary of installing unfamiliar software.
In many institutional situations the option of installing a plug-in to view a DjVu file is no longer possible for an end-user; the software would have to be approved and installed by IT staff. As a consequence too few people have been exposed to DjVu.
I think that JH's suggestions above show a viable way to mitigate these barriers. The academic community will embrace this format if there is sufficient warranty of it not suddenly disappearing or becoming prohibitively expensive and this will ensure that the machines used by students and faculty will be equipped to view these files, as they are in our library.
I still observe the dropped-jaw when someone cognizant of the significance of file-size first sees a properly encoded DjVu image. Former students will employ this technology within the commercial sector if they become familiar with it.
I have recently been creating a demo web app that serves up DjVu docs. People around me are impressed. If we market what I've been working on and it takes off, we're estimating we'll be converting millions of documents. So I hope DjVu is here to stay.