May 16, 2014

Maxim - Before It All Started

The story of Maxim Integrated Products, or Maxim Integrated, or just Maxim, is fascinating. They built Lucite monuments in tribute to Jean Hoerni and Bob Widlar in front of their headquarters in Sunnyvale. Neither Hoerni nor Widlar were employed at Maxim but they are very tightly connected to the story. In one sense, without Hoerni's planar process, modern integrated circuits would have taken a different path or at least been delayed in time. And without Widlar's sheer genius, analog circuits would also have taken a different path or at least been delayed in time. But Hoerni left Fairchild to eventually found Intersil and it is from Intersil that Maxim directly spun-off. And Widlar's first product manager was Jack Gifford (hand-picked by Widlar, according to Gifford), who founded Maxim and was its CEO until his retirement in 2007. But there is so much more to the story. So let’s go back to before it all started.


Jean Hoerni was born in Geneva, and earned a B.S. and his first Ph.D. in physics from the University of Geneva. After earning his second Ph.D. from the University of Cambridge in 1952, he became a research fellow with Linus Pauling at the California Institute of Technology. William Shockley recruited him to join Shockley Semiconductor Laboratories and assigned him to do theoretical calculations of diffusion rates. As one of the “traitorous eight”, he left Shockley Labs and co-founded Fairchild Semiconductor where he devised the "planar process" which made integrated circuits possible. In 1961 he left Fairchild to become vice president of Amelco (the semiconductor division of Teledyne, then TelCom Semiconductor and ultimately Microchip – a story for another day). Hoerni was a consultant for Union Carbide, Hughes Aircraft and Fujitsu before he founded Intersil in 1967.


Bob Widlar grew up in Cleveland and enrolled at the University of Colorado at Boulder. In February 1958 Widlar joined the United States Air Force. He instructed servicemen in electronic equipment and semiconductor devices and wrote the textbook for the course (concurrently). In 1961 he joined the Ball Brothers Research Corporation in Boulder to develop analog and digital equipment for NASA. He simultaneously continued studies at the University of Colorado and graduated in 1963.

His work at Ball Brothers brought Widlar in contact with Jean Hoerni and Sheldon Roberts at Fairchild Semiconductor, regarding radiation hardened transistors. In 1963, Jerry Sanders, at the time a Fairchild salesman but later CEO of AMD, recruited him to join Fairchild. The interview with Fairchild research and development (R&D) manager Heinz Ruegg didn’t go well; Widlar told him what he thought about Fairchild's analog circuits: "what they are doing is bullshit". Fortunately, the Applications Engineering division head, John Hulme, hired Widlar despite objections from the first round interviewers. Widlar's first assignment at Fairchild targeted IC reliability which connected him with process engineer David Talbert.

Fairchild engineers before Widlar designed analog ICs like conventional circuits built with discrete transistors, capacitors and resistors. This was impractical using Hoerni's planar process due to component values and matching requirements so they used thin film hybrids which dramatically limited their reliability and manufacturability. Working in secrecy with Talbert, Widlar soon grasped the benefits and drawbacks of the planar process: it permitted matched performance of all components at all temperatures, but these components possessed parasitic capacitance not present in discrete parts, and the process ruled out the use of large-value resistors and capacitors. Armed with this insight and Hung-Chang Lin's theory of compensated devices (look for references to H.C. Lin in Widlar's early application notes), he designed the μA702 – the industry's first true linear integrated circuit and first monolithic operational amplifier. Widlar did not consider the μA702 prototype good enough for production, but Fairchild decided otherwise and rushed the chip into production in October 1964. Widlar agreed only if he could select the product manager – enter Jack Gifford.


Born in 1941, Gifford grew up in Torrance, CA. He was passionate about baseball and played competitively throughout his school years, including UCLA (the University of California, Los Angeles) where he graduated with a BSEE in 1963. Gifford started his career at Electronic Specialties as a design engineer. At the age of 24, Fairchild Semiconductor recruited him, but insisted on starting him out as a salesman calling on Hughes Aircraft.

Gifford recalled, “I was a salesman, a year out of college in L.A. and I was one of the few guys, myself and I think Vic Grinich, in the company that had any real understanding of gain and phase relationships and how amplifiers worked and things like that. We had a design background and Bob (Widlar) recognized that through dealing with me as a salesman. Then as he and Dave Talbert developed the 709 (probably µA702), they bootlegged samples out to myself and Floyd Kvamme and asked us to see if customers were interested in it as it was a very complicated circuit. In fact no one, maybe with the exception of Dave Fullagar and Bob Widlar, knew exactly what the real circuit looked like. People had the schematic explained to them by myself, Widlar, Mike Markkula and others. That circuit was not the actual circuit; there was another circuit that no one ever saw. We became stars explaining the circuit that was on the diagram. As we sold this idea and explained this circuit, there became a huge interest in the product.”
“So Tom Bay and Bob Noyce went to Bob Widlar, who had defined and designed this circuit all on his own without any sponsorship, and Bob (Noyce) said, ‘you’re gonna be happy to hear that we’re gonna make a product line out of your products’. Bob (Widlar) just looked at 'em both, and remember this is a junior engineer, I mean he’s got a BS degree like I did, he was an applications engineer, shouldn't have known anything about device physics, he taught himself all of that and he says ‘The hell you are.’ He’s talking to Tom Bay and Bob Noyce, two of the most sophisticated, impressive people that I've ever met and he’s telling 'em ‘get screwed, you’re not gonna do this.’ The next day they asked ‘Why not?’ And he said ‘Well first of all you guys don’t know what you’re talking about, you don’t know what the circuit is, you don’t know how it works, and furthermore no one in the company knows how it works and is used. I’m not gonna let you ruin my reputation.’”
“He says ‘Well I want some people competent communicating and dealing with my products.’ They said ‘Okay well we’ll put our best product managers in front of you and you can pick one.’ They ran a bunch of guys through him and he didn't like any of 'em and they finally said ‘Who is it that you would like?’ ‘Well there’s this kid down in LA, Gifford.’ They didn't even know who I was and so I get a call the next day from Bob Graham who was Director of Product Marketing and he says ‘Jack come up and talk about being a product manager for the linear circuit business, we’re looking for candidates.’ I come up and within ten minutes I’m hired, you know, the interview lasted 7 minutes and I go home and I tell my wife I just got promoted, I don’t know why, I have no idea, from being a salesman to running this product line.”“Within a period of a couple of years we owned 80% of the market. The design talent at Fairchild was unbelievable with Bob Widlar, Dave Fullagar, Jim Giles, Darryl Lieux, and Dave Talbert as a process developer.”
“I could tell you the 709 story too about shutting half the world down. If it wasn't for Fullagar's 741, we probably wouldn't have survived as an analog company.”
Widlar and Talbert went to Molectro Science Corporation that was acquired shortly thereafter by Charlie Sporck and National Semiconductor. Gifford took over running the whole analog research and development department. But who would he rely on after Widlar left? – enter Dave Fullagar.


Dave Fullagar described how he started his career.

“Until I was 12 years old, we lived on the moors in the north of England in a house with no electricity, so a crystal set was my only option. In 1954, I read an article entitled How to build a radio in a flashlight. It used something called a transistor – a Mullard OC71. I went down to the local radio store to buy one. ‘Never ’eard of a transistor, boy. Don’t know naught about that,’ said the proprietor in a broad Yorkshire accent.”
“My first job was with Ferranti (a leading UK defense contractor) in Edinburgh, Scotland. I worked on a terrain-following radar for a bomber that was supposed to fly to the Soviet Union at treetop height. They used to fly the engineers over the highlands of Scotland – at 300 feet – to validate the radar’s terrain-following ability. It was a good incentive to get the design right. Ferranti's offer was the best paying – at about $2200 a year; no, I didn't omit a zero – of the four or five jobs offered.”“In the 1960s, there were two principal routes to the United States for Brits. One was via Canada; the other, through Transitron in Massachusetts. My colleague and fellow Cambridge graduate Wadie Khadder had joined Transitron a year earlier, so the transition was an easy one. It soon became apparent that Transitron was in decline. So both Wadie and I moved to Fairchild in early 1966.”
“My assigned task when I joined Fairchild R&D in 1966 was to design the successor to the μA709. The target specification I was given by marketing was of the ‘let’s-improve-all-the-key-specs-by-50%’ variety. However the biggest problems with the 709 were its idiosyncrasies, not its specifications: It was tricky to stabilize, there was no short-circuit protection, and it would latch-up and self-destruct in nanoseconds. National’s LM101, which Widlar designed, addressed many of the user-friendliness issues but still required external compensation and had a kludgy front-end bias scheme. Widlar must have come to the same conclusion, because he later redesigned it as the LM101A with a much-improved front end.”“I proposed the internally compensated μA741 in mid 1967 to Garth Wilson and Marv Rudin, who ran the Linear R&D Group (and later founders of Precision Monolithics Inc.). Next thing I knew, I was sitting in Gordon Moore’s office. He asked me if I’d mind moving to Mountain View because that would expedite the introduction of the part, which occurred in May 1968.”
“It was exciting, it was just a creative period when Fairchild was absolutely king of the heap and thanks to Jean Hoerni's planar process, it had the best processing in the industry, the highest performance transistors, a very successful logic family. In early 1966 a processing R&D group under Andy Grove was writing the book, it turned out literally, on semiconductor processing and the emerging MOS technology. A prolific consumer group and absolutely number one in linear circuits, in fact honestly there was nobody else in the business at that time. (Widlar) and Talbert formed this incredible duo and with almost no official sponsorship from Fairchild created a linear family, the 702, 709, 710 and 711. Those products went on to form the basis for the Fairchild linear products and I think all of us in the design community pay homage to that effort because it was so incredible.”
“By 1968 there were a host of second generation products under development, Dick Lane’s 722 D to A, Colin Barry’s 715 high speed amplifier, Darryl Lieux’s 723 voltage regulator, Bill O’Neil’s 733 wide band amplifier, George Erdi’s 725 we’ve already mentioned my 741 and a host of other products. Fairchild seemed to rule the world with National really being the only competition in sight.”
“Then something happened and it all seemed to unravel. Between the fall of ’68 and the spring of ’69 Jack together with Jim Giles, Larry Stenger and Frank Botte went off to form AMD. Marv Rudin, Garth Wilson, George Erdi went off to start PMI. I joined Jean Hoerni, Don Rogers, and Murray Siegel at Intersil shortly thereafter joined by Bill O’Neil. Dave Bingham went to Cermatek, Colin Barry went to Signetics, Len Brown went to Motorola, and Mike Markkula joined Intel (the latter two from the marketing group).”
“Speaking for myself, I don’t think it’s because I thought that I was going to get rich by going to Intersil. Frankly I didn't even know what an IPO was, you know, I didn't know what a stock option was in any real sense of the word and it wasn't because I had in any inkling that Fairchild was about to implode. It seemed like I was leaving at the crest of the wave frankly. Neither is it really fair to blame it all on the cultural change brought on by the arrival of the guys from Motorola, although it did seem like a management style change which maybe took some of the fun out of Fairchild as I perceived it. But I think the overriding factor for me anyway was that I was looking maybe for more of a challenge, I wanted to start my own group, have more say in the way the company ran and Intersil at that time offered that opportunity. My office was two doors down from Jean Hoerni's, company strategies were discussed around the coffee table and by comparison, Fairchild seemed kind of ponderous.”
So let’s get back to Gifford for a moment. He had the inspiration that a pure analog company could do really well. But his first attempt was forced in a different direction. He sought funding for a startup company, Advanced Micro Devices (AMD, perhaps you've heard of it?), but was repeatedly turned down because the financial backers wanted someone with a little more experience. Enter Jerry Sanders. Together they finally obtained funding in 1969. Sander’s version goes like this: “First a young man named Jack Gifford, who I'd worked with, you know, at Fairchild, who had an idea to start a company to make linear integrated circuits? And that was interesting to me, but frankly linear integrated circuits in my view were a niche opportunity and although the analogue world will always be with us, because that's what the real world is, I just didn't see that as an exciting thing whereas I was more interested in the digital world where you could build more and more complex things.” Sanders was CEO and Gifford was chief financial officer. Sanders was not popular with the engineers and Gifford had to tell Sanders to change his tyrannical ways. Sanders threatened that any staff members supporting Gifford would be let go. Gifford was asked to leave in 1971. Gifford says only that he had a "falling out" with Sanders. Don Valentine speculated that the company simply wasn't big enough for two such strong personalities. "The diameter of a spotlight only fits so many people," says Valentine. "And Jerry was an individual who always liked to be in the spotlight."

Meanwhile, Fullagar had gone to Intersil as the first analog designer working for Jean Hoerni. “I think Intersil lured me, but it wasn't financial. I didn’t really know what a stock option was and couldn’t have distinguished between an IPO and a UFO. Intersil offered a chance to create my own analog-design group with a clean slate and to work with Dr. Hoerni, one of the true giants of the semiconductor industry.” He added, “Intersil was primarily a discrete transistor and FET company prior to my joining. FETs were Jean Hoerni's first love. I was the first analog IC designer and about employee #50. I worked initially for John Hall, who later founded Micropower Systems (and Linear Integrated Systems). I brought Bill O’Neil into the group; he had designed the µA733 at Fairchild.”

Fullagar knew Gifford from Fairchild. “I met Jack shortly after joining Fairchild in early 1966. He was in charge of Linear Marketing and had recruited a team of young engineers including Mike Markkula, Mike Scott (both of whom became Apple CEO at different times), Len Brown, Jerry Zis, and others. We were all in our mid twenties at the time and used to socialize together at the Wagon Wheel and other places after work. At that time I did not see any of the hard-driving potential CEO material in Jack; we were just a bunch of kids having a good time.”

For a while, Gifford turned to farming, building a small fruit orchard near Sacramento into a successful 2,000-acre tomato producer. Fullagar was unsatisfied, “When I went to Intersil, it was frustrating to be designing products that I felt weren’t selling because of inept marketing.” In 1971, Jean Hoerni, head of Intersil, recruited Gifford to help them enter the analog data-acquisition business. In exchange, he would help Gifford raise money to start his own business. For several years he worked there part time, while continuing his farming and other business interests. Eventually it became full-time. After a succession of somewhat inept CEOs, Orie Hoch (ex-Litton CEO) was brought in by the Intersil board and did a great job of straightening out the company then sold it to GE. Once the sale was completed, Hoch made Gifford CEO and returned to Litton.

Gifford agreed but only if he was given stock options and if Intersil could be an independent GE subsidiary. Jack Welch agreed and bragged about his latest acquisition. Contrary to stories at the time, Gifford and Welch were kindred spirits and got along well together. However Intersil was put in the light bulb division of GE under a sector VP called Jim Baker. He and Gifford did not get along. There was a big showdown between the two of them over employee compensation (Baker didn't think stock options were necessary since his people in Schenectady didn't have them).

According to Gifford, this is what happened:

“Well, finally after about six months of this, he starts to hear from his other general managers. Hey, this isn't fair. I mean, I've worked for GE all my life. I don't have stock. What's going on, I mean, how could he do this? And he had a mutiny on his hands. And so Jack (Welch) started talking to me, you know, and every once in a while he'd say, you know, ‘God, you know, boy the stock options are causing me a problem, you know.’ Then it got, ‘Do you really need those? Can't we do something else?’ ‘Jack, that was the deal,’ I says, and ‘I don't need them. I mean, but Intersil needs them. I can't, by the way, that was your idea.’ He says, ‘I know. I know.’ So anyway, then we're back at one of these parties again; about a year later. I'm standing around talking to ten or fifteen of these guys and Jack comes across the room and he's got ten or fifteen guys following him. I get there late. So he's already had a couple of drinks, and he comes swaggering into my group and, you know, how the hell are you. We start talking. No sooner does the greetings stop and he says, ‘God damn it. Why the fuck do you guys have to have those stock options? They’re, you know, jeez.’ And he goes on and on and on in front of fifteen guys. And all of a sudden he's now, you know, coming like it's my fault, like I'm the problem, right? And he's, say he's had a couple of drinks. And finally he starts, ‘Well, what do you think?’ And I said, and I was going to react and I says, I was so mad and I just said, ‘Fuck you.’ And I just turned around and walked out. And man, I heard that rippled throughout the company. I mean, you told that to Jack? Yeah, but, you know, yeah, I did. So about a week and a half later I get a call from him to come back to Fairfield, and he says, ‘Well, I gotta talk to you.’ What are you gonna do? He says, ‘Well, I gotta fire you.’ So he fired me.”
“Yeah, and what happened is that's what caused Maxim. So when he changed it (the stock options), those guys came to me and said, you know, Jack, you gotta do something, you know. Our careers, he's destroyed our careers, you know. And I was forced to do Maxim. I was, I mean, I caused the problem, so...”
The idea Jack had about Maxim was to build the best analog company in the world that was owned by employees. He felt that analog had always been secondary; generating profit to fund digital. He felt he was the best in the world. But he says “I was not motivated to do it. Frankly, what motivated me was this responsibility to these other guys.”
Photo of the original Maxim founders, taken at Gifford's house the day most of them quit Intersil after funding was secured.  From left to right: Bev Fuller, Rich Hood, Dave Fullagar, Fred Beck, Roger Fuller, Sam Ochi, Dave Bingham, Lee Evans, Jack Gifford, Steve Combs. 

According to Pirooz Parvarandeh, “The first person to come out of the company was Jack, followed by Fred (Beck) and Dave (Fullagar) and a few other people. Their concept was to start a business around what they were doing at Intersil — analog semiconductors. So their first priority was really to secure some venture capital. To do this, they created a business plan. The business plan, a three-page business plan, is actually framed and posted at our headquarters.”

“The story goes that Jack, Fred, and Dave worked on this plan on a napkin. There was a debate as to whether a three-page business plan was long enough. Or should it be 20 pages or 50 pages? They were debating whether they needed to be much more sophisticated about the business plan. As it turns out, Jack convinced his colleagues that three pages was enough. It is a very high-level view of what they wanted to do.”
“Obviously, it didn't have a lot of detail. It had some calculations of what the market size would be and what the opportunity would be. It needs to be said that Jack was a very dynamic and energetic individual. They were able to raise $10 million from a number of venture capital firms.”
Wikipedia says the business plan was two pages and the capital was $9 million. The founding team included Gifford; Fred Beck, an IC sales and distribution pioneer; Dave Bingham, General Electric’s Scientist of the Year in 1982; Steve Combs, a pioneer in wafer technologies and manufacturing; Lee Evans, also a pioneer in CMOS analog microchip design and General Electric’s Scientist of the Year in 1982; Dave Fullagar, inventor of the first internally compensated operational amplifier circuit; Roger Fuller, yet another pioneer in CMOS microchip design; Rich Hood, development director for some of the first microprocessor-controlled semiconductor test systems; and Dick Wilenken, who is acknowledged as the father of key analog switch and multiplexer technologies. In the first year, the company developed 24 second source products before turning to predominantly proprietary products.

Fullagar mentioned another founder, Sam Ochi. “We hired Sam from Teledyne so no one could accuse us of being an entirely Intersil spinoff, but that didn't prevent GE from suing us.”

Dave Fullagar said, “In the early days, five or six of us (including Jack) would meet every Tuesday morning for a product planning meeting at the Peppermill restaurant on De Anza Blvd. Sometimes Jack would forget that we were in a public place and the four letter words would start flying: the restaurant manager moved us into the far back corner! Most of the ideas came from designers (especially Dave Bingham) and Charlie Allen (applications). All the early successes (RS-232 circuits, microprocessor supervisors, etc) evolved from these meetings. Eventually we outgrew this modus operandi and the process became more formalized (but not as much fun!).”

“The really novel products come about in one of two ways. One is when an engineer figures out a totally radical approach to meet a functional need. For example, Widlar's bandgap reference was sheer genius. We all knew that VBE had a negative temperature coefficient and delta VBE had a positive temperature coefficient, but Widlar figured out how to put the two together to produce a temperature-stable reference.”
“The other way is when a good applications engineer works hand in hand with a talented designer. Maxim’s RS-232 products came about in this manner. Charlie Allen, the best applications engineer I have ever encountered, recognized that lots of digital equipment had ±12V supplies whose sole purpose was to power RS-232 drivers. RS-232 drivers didn't fit into any of Maxim’s analog-product lines, but he persisted. He corralled Dave Bingham, one of Maxim’s most creative design engineers, into taking a look at it. Bingham felt sure he could put charge pumps on the same IC as the RS-232 drivers — no mean feat — thereby eliminating the need for the ±12V supply. Thus was born one of the company’s most successful product lines.”
Parvarandeh recalled, “When they started the company, there were some well-known names in the industry. Analog Devices, for one, has a much longer history than us. They were in the market well ahead of us, as well as Texas Instruments and other players. One of the first priorities of the new company was to make sure that they generated some revenue. Also, there was a spirit of inventiveness in the company. You know, you try to come up with innovations that are really world class. It was clearly part of the DNA of the company to try to distinguish itself.

Fullagar echoed the same idea, “The original business plan called for the speedy introduction of 14 second-source parts to generate quick cash flow, followed by proprietary parts. Generating positive cash flow in a start-up is key: If you have to go back to the venture capitalists for an unscheduled round of financing, you get taken to the cleaners.”

Again from Parvarandeh: “To the extent that we could come up with enhancements that we could make, we would make them. In fact, I remember — I joined the company in 1987 — that one of the things I was striving for was, ‘Okay, if I’m doing a second source of a product, is there something else that I could do that would benefit our customers?’”

“So, the first order of business was, ‘Let’s try to make sure that we capture some existing sockets, but also let’s see if we can provide additional benefits.’ That was the strategy and, of course, for a startup company to try to create or to leverage existing business was the right strategy. But as I said, the company had in its DNA this notion that we don’t want to be just a second source company. We want to be a company that creates valuable product differentiation.”

Maxim has always been an aggressive company. Most of their growth was organic, but some resulted from acquisitions. Key acquisitions include (or at least the ones that I think are key):

  • 1994: Tektronix Semiconductor Division, for its high-speed bipolar processes, wireless RF and fiber-optic products
  • 2001: Dallas Semiconductor
  • 2007: Vitesse Semiconductor’s Storage Products Division
  • 2010: Teridian Semiconductor Corporation, supplying systems on a chip (SoC) for the smart meter market
  • 2010: Phyworks, a supplier of optical transceiver chips for broadband communications
  • 2011: SensorDynamics, for proprietary sensor and microelectromechanical (MEMS) solutions
  • 2012: Genasic Design Systems Ltd., a fabless RF chip company that makes chips for LTE applications
  • 2013: Volterra Semiconductor

I enjoyed creating this particular post. I’m thankful for the interview transcripts that helped my research. And I am very thankful to Dave Fullagar for contributing to this and clearing up some things. In college, I was taught about operational amplifiers using the 741. It’s surreal to be able to simply contact him and ask a few questions. Fullagar is famous for the 741, of course, but that was only one of many things he accomplished in his career. In another article, Fullagar reflected on his career, “I thrive in a small-company atmosphere, so I would have to say the early days at both Intersil and Maxim were the most enjoyable. The lab work was always especially rewarding. Checking out a new breadboard with a cup of coffee warming on top of a big Tektronix 545 oscilloscope — that was real engineering! Sitting in front of a screen running simulations just isn’t the same. I feel proudest of the design team I recruited. They were — and still are — some of the finest people in the industry, having designed an incredible number of innovative products, as well as serving in senior-management positions, including current President and Chief Executive Officer Tunc Doluca and several of the vice presidents.”

In closing, I’d like to leave you with this opening from Maxim application note, AN1795:

It's a jungle out there.

A small tribe, in the dense wilderness, is much sought after by head hunters from the surrounding plains. Known throughout the land for their esoteric expertise, this is the tribe of the Analog Engineers, who live in the farthest regions of the left half Plains, past the jungles of Laplace.

The guru of analog engineers is the Analog Filter Designer, who sits on the throne of his kingdom and imparts wisdom. You never get to see him, even with an appointment, and you call him "Sir."

The countless pages of equations found in most books on filter design can frighten small dogs, and digital designers. This article clears a path through the brush for the practical engineer and unravels the mystery of filter design, enabling you to design continuous-time analog filters quickly and with a minimum of mathematics.

Analog electronics has two distinct sides: the theory taught by academic institutions (equations of stability, phase-shift calculations, etc.), and the practical side familiar to most engineers (avoid oscillation by tweaking the gain with a capacitor, etc.).

Yes, filter design may be closer to mystery than mathematics, but that whole passage applies well to most disciplines of analog design.



Fairchild Oral History – Fullagar, Gifford, Wilson 

“Jack Gifford: Baseball’s Loss Was The World’s Gain”, Doris Kilbane, Electronic Design, Dec. 7, 2009

Silicon Genesis – interview with Jack Gifford

Interview: Maxim Integrated CTO Pirooz Parvarandeh Discusses Analog/Digital Integration, Joe Desposito, Electronic Design Aug 12, 2013

No comments:

Post a Comment